admin.html · 17 screens · App mode (admin trust+ chrome)

Admin operations

A1 platform operator + A2 analyst + A4 moderator surfaces. Plus ops dashboards (F.X*).

📋 SCENARIO Admin lead (A1) opens the executive dashboard at the start of shift to scan platform health, current Wedge phase, and any cross-team escalations from A2/A3/A4. They drill into the highest-severity SLO breach or compliance flag and route ownership to the right pod.
19Admin overview (executive dashboard)
A1 · /dashboard (role: admin)
⚡ PHASE 0.α
Wedge F · Quotation→Commerce · v1 launched 2026-05-04 · F-acceptance criteria 4 / 7 met
View service catalog →

Platform overview

Last refreshed 2m ago · Auto-refresh on · Trust + ops triage
SA
A1.sarah_admin
Switch role →
Active providers
2,447
+18 this week
Active patients
14,892
+847 / 7d
Quotations / day
1,247
P95: 1,420
GMV / 30d
$8.4M
↑ 12%
Revenue at risk
$12.4k
1.5% of GMV · 3 open disputes

Provider verification pipeline

Click → verification queue ›
47
Applied
39
Auto-pulled
8
A1 review
31
Approved
7
Pending
2
Critical >7d
Trust primitive (NEW V3 verification scheme) · 2 stalled providers exceeded the 7d SLA.

Patient escalations

P1 → A1 from p43 chat threads
4
Open
2
Awaiting A1
Avg response3.2h
SLAwithin 4h target
🚩 Fraud signals · F.X4 surfaceAuto-detected 14m ago
3 deposit-velocity anomalies 1 IP-cluster pattern · 14 sessions 2 cross-border quotation-spam patterns

SLO health

Drill into SLO dashboard →
Quotation submit p95187ms · within budget
Notification dispatch p951.2s · within budget
Token validation p9552ms · 4% over budget
Cold-start TTHW218s · within budget

Open incidents

2 active
P2
Tolgee translation API latency elevated
Started 18m ago · A2 on-call
P3
Discovery embedding job behind
Started 1h ago · backlog 4,128 items
📋 SCENARIO Admin reviewer (A1) works the verification grant queue, validating provider credentials, business licenses, and medical-board registration before granting Verified status. SLA 48h; ambiguous credentials escalate to senior A1 with rubric notes attached.
33Verification grant queue (A1)
/dashboard/verification-queue

Verification queue

Trust primitive · NEW V3 verification scheme replaces ArokaStars · binary verified + Trust+ premium tier
7 pending2 critical >7d
All · 47 Critical · 2 A1 review · 8 Docs needed · 12 Approved (30d) · 31 Phase 0.β will add bronze · silver · gold tiered verification
Selected (0)
Provider
Country
Submitted
Time in queue
Auto-pull
Assignee
Status
Bumrungrad International
Re-verification (license)
Thailand
2026-04-18
21d · ok
JCI ✓ MCT ✓
A1.sarah_admin
APPROVED
›
MedPark Hospital
Initial application
Thailand
2026-05-02
7d · breach soon
JCI ✓
A1.sarah_admin
A1 REVIEW
›
Krabi International Hospital
Initial application
Thailand
2026-05-04
5d · ok
JCI pending
— unassigned
DOCS NEEDED
›
Bangkok Hospital Phuket
Re-verification (2yr default)
Thailand
2026-05-05
4d · ok
JCI ✓
A1.anita_v
A1 REVIEW
›
Khon Kaen Ram Hospital
Initial application · cosmetic
Thailand
2026-04-28
9d · CRITICAL
2 of 4
A1.sarah_admin
CRITICAL
›
Bangkok Hospital Hat Yai
Initial application · plastic
Thailand
2026-04-29
8d · CRITICAL
1 of 4
— unassigned
CRITICAL
›
Open · MedPark Hospital · evidence pack
Initial application · Thailand · A1.sarah_admin reviewing · trace_id 4bf92f3577
A1 review
Credentials · auto-pulled
JCI accreditation · MJ-2024-03898% confidence
Thai MoPH license · ABC-01494% confidence
HA Thailand certificationmanual verify
Tax ID · validatedtreasury match
Practitioner roster · 47 doctors
Spot-checked 6 of 47 against local registryall matched
Outcome history sample (18 cases)
Discharge w/in 5 days · readmission < 30d17 / 18 normal
Complaint history (last 24mo)
Filed complaints · resolved2 · both resolved
Decision · A1 rationale required
Every action signs an audit-log entry (F.X2). Phase 0.β: bronze/silver/gold tiers.
📋 SCENARIO A3 support operator searches the user roster to handle a patient or provider help ticket — resetting MFA, reassigning roles, or merging duplicate accounts. Destructive actions (suspend, delete) require dual A1 sign-off and are logged to the immutable audit tail.
21User management
/dashboard/users

Users

RBAC + KYC + impersonation log · every action signs F.X2 audit chain
Active 24h
1,847
↑ 3% wow
KYC pending
14
avg wait 2.4d
MFA disabled
87
F lock breach
Suspended
9
5 fraud · 4 violation
All · 17,386 Patient · 14,892 Provider · 2,447 Admin · 47 Suspended · 9
Selected (0)
User
Type
Country
Email
Created
MFA
Status
Sarah Khan
Patient
US
sarah.khan@…
2026-04-22
TOTP
ACTIVE
⋯
Marco Donatello
Provider
TH
marco@bumrungrad…
2025-11-03
YubiKey
ACTIVE
⋯
James Liu
Patient
AU
james.liu@…
2026-05-01
none
UNVERIFIED
⋯
Somchai Wong
Provider
TH
somchai@samitivej…
2025-08-12
TOTP
ACTIVE
⋯
AnonymousUser_42
Patient
—
— private —
2026-03-04
none
SUSPENDED
⋯
Invite admin · modal previewTriggered by + Invite admin
Email
Role
Sub-tier (if A4)
Invite consumes a one-time enrollment URL (24h expiry, MFA setup forced).
📋 SCENARIO A1 leadership reviews GMV, take-rate, subscription MRR, and ad-revenue mix to track Wedge F monetisation health. They flag take-rate compression by archetype to A2 for cohort analysis and brief Finance on weekly run-rate.
22Revenue analytics
/dashboard/analytics/revenue

Revenue · last 30 days

A2 + A1 view · Insight primitive (V2.5 + commerce events) · GMV / take rate / MRR · Wedge F commerce visibility
🚩 Cardiology bookings ↓ 12% wow
Auto-flagged · likely cause: Phuket International Hospital re-verification queue + 2-week JCI audit window
Investigate →
GMV
$8.4M
↑ 12%
Platform fee
$420K
5% take
Subscription MRR
$84K
+9 Trust+
Avg booking
$2,840
↓ 4%
GMV trend · 12 months↑ 12% yoy
Jun '25May '26
Take rate · 12 months5.0% steady
Target 5.0%Current 5.0%
Bookings by category · 30d
Dent · Cosm · Card · Orth · Fert
Per-channel attributionClick-attributed bookings, last 30d
Organic 62% SEO 18% Partner 12% Direct 8% Wedge I will add ad-attribution column
MRR split by tier
Free · 2,396 providers$0
Trust+ Premium · 105 providers$84,000
Wedge I tiers · inactive— grayed —
Subscription primitive ships at F · only Free + Trust+ Premium configured

Per-region breakdown · top destinations

Destination
Bookings
GMV
Avg booking
% of total
Bangkok, Thailand
847
$2.84M
$3,355
34%
Chiang Mai, Thailand
524
$1.62M
$3,090
19%
Phuket, Thailand
412
$1.42M
$3,447
17%
Pattaya, Thailand
214
$840K
$3,925
10%
Hua Hin, Thailand
182
$580K
$3,187
7%
Krabi, Thailand
98
$320K
$3,265
4%

Procedure category · 30d

Procedure category
Bookings
GMV
Avg value
Conv rate
Cosmetic
847
$3.2M
$3,780
8.4%
Dental
1,124
$2.0M
$1,780
9.7%
Cardiology
182
$2.4M
$13,200
4.2%
Orthopedic
214
$840K
$3,930
6.1%
Fertility
98
$590K
$6,020
5.8%
Scheduled reports
Weekly GMV digest
Mondays · 09:00 ICT · 8 recipients (board + A1/A2)
ENABLED
Monthly take-rate report
1st of month · CSV + PDF · 4 recipients
ENABLED
📋 SCENARIO Admin oversight (A1) audits the sponsored-listing auction for collusion, reserve-price violations, and tier-eligibility drift. Anomalous bid patterns trigger a freeze on the affected slot and an escalation to A2 to inspect the auction model.
62Sponsored auction dashboard (A1 oversight)
/dashboard/ads/auction
✓
Wedge I live · Advertisement primitive online · 30% sponsored cap enforced · A4 ad-policy review chain active

Sponsored auction · 30-day overview

Growth primitive · auction analytics · medical-ad policy review chain
Active bids
847
↑ 14% vs prior 30d
Avg CPM
$8.40
range $4–$24
Sponsored cap usage
22%
Below 30% hard cap
Per results-page · advertisement.md L156
Ad revenue · 30d
$84k
projected $112k Q3
Inventory cap policyhealthy
22% current usage8% available30% hard cap
Hard cap protects discovery quality · cannot be exceeded without A1 + product-lead approval.
Revenue share
ArokaGO 30% · platform feeProvider 70%
Disclosure surfaced in provider auction settings + ToS.
Per-tier bid behaviour
Free providersfull bid
Pro10% bid discount
Premium15% bid discount
Enterprise20% discount + reserved slots
Tier-aware pricing
All 4 I-tiers configured · ranking-boost multipliers per wedge spec.
Freebase · 1.0×
Pro1.1×
Premium1.25×
Enterprise1.5×
⚠ Policy violation queue · 3 ad copy items in A4 ad-policy review
1 medical-claim flagged ("guaranteed cure") · 1 unverified procedure efficacy · 1 missing destination disclosure
Open A4 ad-policy queue → #a65

Top winning bids

Provider
Slot
Tier
Bid
Win rate
Spent / 30d
A4 creative
Bumrungrad
Bangkok · Dental implant
Premium
$12 CPM
68%
$2,840
✓ approved
›
BNH Hospital
Bangkok · Rhinoplasty
Enterprise
$10 CPM
72%
$2,140
⚠ in review
›
Chiang Mai Ram
Chiang Mai · Knee replace
Pro
$8 CPM
82%
$1,920
✓ approved
›
Auction-collusion monitor7d · 0 alerts
A2-classifier flags suspicious bid coordination patterns (per wedge spec risk row 5 + open question 13).
Synchronised bid clusters0
Round-robin win patterns0
Cross-provider IP overlap0
Last classifier run14 min ago
Per-city × category auction depth6 × 6 grid
Cell = active-bidder count + clearing-CPM. Click cell → drill into city-category auction history.
Dental
Cardio
Ortho
Cosmetic
Fertility
Wellness
Bangkok
14 / $11
8 / $9
6 / $7
11 / $13
4 / $6
5 / $5
Chiang Mai
7 / $8
5 / $7
6 / $7
12 / $14
2 / —
4 / $5
Phuket
5 / $6
7 / $8
9 / $9
2 / —
8 / $9
1 / —
Pattaya
9 / $10
6 / $8
5 / $7
7 / $11
6 / $9
8 / $7
Hua Hin
8 / $9
2 / —
4 / $6
13 / $15
5 / $7
6 / $6
Krabi
5 / $6
4 / $6
2 / —
6 / $8
1 / —
7 / $6
Dashed cells = below auction-depth threshold · reserve-price floor recommended.
Wedge I monetization surfaces
Inventory definitions · A4 ad-policy review · subscription analytics.
📋 SCENARIO A4 moderator works the consolidated content queue — flagged listings, reviews, community posts, and chat reports. They apply policy actions (approve / hide / take-down / strike), and severe violations are forwarded to A1 for account-level enforcement.
20Moderation queue
A4 · /dashboard/moderation

Moderation queue

A4 multi-sub-tier surface · Trust + Compliance + Conversation primitives · auto-classifier first-pass + A4 review + A1 escalation
Critical 2 High 7 Medium 23 Low 14 A4.translation 4 A4.editorial 6 A4.community 23
Selected (0)
Severity
Type
Content
AI confidence
Reporter
Assignee
Submitted
CRITICAL
Medical accuracy
"Cancer can be cured by traditional…"
96% violation
Auto-classifier
A4.community
3m ago
›
CRITICAL
PHI leak
Provider posted patient image w/o consent
99% PHI
Sarah K. (patient)
A1.escalated
14m ago
›
HIGH
Spam
Repeated cross-thread provider promotion
88% spam
3 reports
A4.community
52m ago
›
A4.TR
Translation review
Article: "What to expect from CABG" — JA review
94% TM
System
A4.translation
2h ago
›
A4.TR
Translation review
Provider profile: BNH Hospital — AR review
91% TM
System
A4.translation
3h ago
›
MEDIUM
Editorial
Article: "Choosing a CABG surgeon abroad"
87% ok
System
A4.editorial
4h ago
›
Open · medical-accuracy violation · CRITICAL
Posted by AnonymousUser_42 in thread "Thai herbal medicine" · 3m ago · classifier: medical-accuracy v2.4
96% confidence
Flagged content
"Cancer can be cured by traditional Thai herbs at this clinic. Skip chemotherapy — these herbs work better and have no side effects. DM me for the clinic name."
Classifier matched: medical-claim-without-evidence + treatment-substitution + off-platform-redirect
Poster history
2 prior strikes6mo on ArokaGO14 posts
A4 decision
Assignee
Decision rationale
📋 SCENARIO A4 editorial moderator reviews provider-authored long-form content (procedure guides, doctor bios, case studies) for medical accuracy, tone, and policy compliance. Inaccurate claims are returned with rubric notes; persistent violations escalate to A1 for credential review.
78Editorial queue (A4)
/dashboard/moderation/editorial

Editorial queue

Articles, photos, provider-authored copy + Wedge I ad copy awaiting A4.editorial review.

📌 At Wedge F · A3 Content Producer is inactive
This queue holds V2.5 inherited submissions only. Wedge G+ activates A3.provider, A3.freelance, A3.patient sub-tiers + per-locale review chain via Tolgee.
Articles 12 Photos 24 Storefront copy 8 Ad copy (Wedge I) 3
"How to choose a CABG surgeon abroad"
Article · A3 author: Dr. Anil K. · 18 min read · Cardiology · v2 vs v1 published 2026-03-12
A4.editorial review
Diff view · split-pane
SOURCE · v1 (published)

For coronary artery bypass graft (CABG) procedures abroad, JCI-accredited Asian centers offer outcomes equivalent to top US/EU centers at 30-50% of the cost. Patient selection is critical.

PROPOSED · v2

For coronary artery bypass graft (CABG) procedures abroad, JCI-accredited Asian centers have shown comparable mortality (1.2% vs 1.4% US benchmark) at 30-50% of the cost (peer-reviewed: Singh et al. 2024). Patient selection is critical.

⚡ AI-assist suggestions (Tolgee)
Simplify title (9 words) Clarify dosage in §4 Add CTA · Schedule consultation
Bumrungrad before-after gallery (4 photos)
Storefront edit · consent verified · Dental implant · single tooth · A3.✗ inactive · provider PR4-authored
before · 1
after · 1
before · 2
after · 2
BNH Hospital · sponsored ad copy "Guaranteed cure for…"
Wedge I · medical-claim flagged by classifier · 99% confidence violation
WEDGE I · ad policy
"Guaranteed cure for chronic back pain at BNH Hospital Bangkok. 100% success rate or your money back."
📋 SCENARIO A4.translation moderator reviews machine-translated medical content in the Tolgee UI, focusing on low-confidence segments and clinically sensitive strings. Approved translations sync to next-intl for production rendering across all 13 locales.
79A4 translation review (Tolgee UI integrated)
/dashboard/moderation/translation

A4.translation review · Japanese

"What to expect from a CABG procedure abroad" — auto-translated by Tolgee, awaiting native review · Wedge K full activation · Tolgee bridge alive at F (13-locale deposit confirmation translation)

⛨ A4.TR REVIEWER
⚠ Clinical-accuracy second pass required
Medical content (CABG · cardiology) requires PR4.Health domain review before publish.
SOURCE · ENGLISH

For coronary artery bypass graft (CABG) procedures abroad, JCI-accredited Asian centers offer outcomes equivalent to top US/EU centers at 30-50% of the cost. Patient selection is critical: stable angina patients with no acute coronary syndrome are best candidates for medical-tourism CABG.

TARGET · 日本語

海外での冠動脈バイパス術 (CABG) では、JCI 認定のアジア医療センターが米国/欧州の最高水準と同等の治療結果を、コストの 30〜50% で提供しています。患者選択が重要で、急性冠症候群のない安定狭心症患者が、医療ツーリズム CABG に最も適しています。

⚡ Tolgee AI translatedConfidence 94%
Tolgee UI · in-context translation editor
ArokaGO glossary · term consistency
Locked terms across this content
CABG冠動脈バイパス術
JCIJCI 認定
stable angina安定狭心症
acute coronary syndrome急性冠症候群
Terms consistent across 142 articles · last updated 2026-04-22
Translation memory · 17 segments
12
Exact match
4
Fuzzy
1
New
Multi-locale fan-out
After JA approval · 12 locales fan out via Tolgee
JA · current KO ZH AR ES PT TH +5 more
Tolgee + locale-specific A4 reviewer required for each locale.
📋 SCENARIO A4 community moderator triages flagged forum posts, comments, and DMs for harassment, medical misinformation, or PHI leakage. They issue warnings, hide content, or escalate to A1 for ban-level enforcement when patterns emerge.
80A4 community moderation
/dashboard/moderation/community

Community moderation

Threads, posts, reviews flagged for review · Trust + Compliance + Conversation primitives

2 critical7 high
CRITICAL · MEDICAL ACCURACY 2 prior strikes 6mo on ArokaGO · 14 posts
"Cancer can be cured by traditional Thai herbs at this clinic…"
Thread reply in "Thai herbal medicine alternatives" · poster: AnonymousUser_42 · 3m ago
SPAM · 3 REPORTS 8 cross-thread posts provider account
"Visit my clinic — best prices guaranteed!"
poster: provider BNH Hospital · marketing intern account · 52m ago
FALSE-REVIEW SUSPECTED flagged by provider
"⭐ 1 star · Bumrungrad scammed me out of $4k…"
Review on Bumrungrad profile · no corresponding Quotation/Booking · poster recently created · 2h ago
Community guidelinesFull →
  1. No medical claims without evidence
  2. No PHI / patient images without consent
  3. No cross-thread provider promotion
  4. No off-platform booking redirects
  5. Reviews require verified booking
  6. Respectful tone · no harassment
Appeal flow
Banned posters can appeal in 7 days via /appeal form. A1 reviews appeals weekly.
Open appeals3
Appeal SLAavg 2.4d
A4.community team
A4.community.alexon duty
A4.community.mayaon duty
A4.community.kimoff · M
📋 SCENARIO A4.live moderator joins a provider's live clinic tour or doctor Q&A as an observer with mute/cut authority. Real-time deepfake detection runs at 99% confidence; cross-language transcripts flagged for A4.translation review on confidence drop.
81Live Q&A moderation (A4.live)✓ Wedge K live · A4.live moderation active + deepfake detection mature
/dashboard/moderation/live

Live · "Knee replacement AMA with Dr. Anil"

Bangkok Hospital · Aug 9, 19:00 UTC · A4.live moderating · Trust + Conversation primitives

● LIVE · 247 viewers
Provider
Bangkok Hospital
Doctor
Dr. Anil K. (PR4)
Viewers
247 live · 12 countries
Queued Q's
18 pending
Flagged
3 · auto-classifier
[ live video stream · Dr. Anil ]
● transcript live-tail · auto-flagged keywords highlighteden → multi-locale subs
19:08:42 · So when we talk about robotic-assisted knee replacement, the key benefit is precision…
19:09:01 · The recovery typically takes six weeks before flying long-haul.
19:09:24 · We don't recommend skipping post-op follow-up just to save costs…
19:09:42 · ⚠ keyword "guaranteed cure" detected · auto-flag for A4.live review
Question queue · 18 pending3 flagged
James L.: How long is the typical recovery before I can fly home?
Sarah K.: Do you do robotic-assisted?
Anonymous: Does anyone want to share my survival story…
⚠ Auto-flagged: spam pattern (94%)
Maria S.: ¿Tienen un cirujano que hable español?
⚡ Tolgee auto-translated: "Do you have a Spanish-speaking surgeon?"
✓ Deepfake checkre-check 30s
Speaker face matches PR4.Health profile
Detector confidence 99% · last frame 19:09:38 · K-spec line 127
A4.translationCross-language
EN primary · TH + ZH live · AR queued
Backup queue 3 items · A4.translation reviewing flagged terms per locale
EN ✓ TH ✓ ZH ✓ AR · 3
Strike escalationUser.spam_strike_count
Strike 1warning
Strike 224h ban
Strike 3permanent ban
AI auto-flag confidence thresholdA2 tunable · trust-and-safety.md line 63
Auto-action
≥ 95%
no human in loop
Human review
60 – 95%
A4.live queue
Ignore
< 60%
log only
[ 0% ────────●────────●──────── 100% ] · auto-action ≥ 95 · review 60 · ignore below
Session archive
After session ends · full transcript + flagged moments + actions taken archived for 7y · medical content retention · /dashboard/moderation/archives
📋 SCENARIO A2 release engineer ramps a Wedge feature flag by archetype + locale, watches the SLO + error-budget impact, and aborts to 0% if budget burn exceeds threshold. Roll-forward decisions require a paired A1 sign-off and are recorded in the audit tail.
82Feature flag dashboard (F.X5)
/dashboard/ops/flags

Feature flags

F.X5 lock · per-tenant / per-region / per-locale scoping · gradual rollout · audit-logged · role-gated
All · 38 Wedge F · 14 Wedge G · 9 Wedge H · 7 Wedge I · 5 Wedge K · 3
Flag
Description
State
Scope
% rolled
Owner
Last changed
quotation.ai_match_v2
Wedge F
AI Quotation matching v2 · embedding-based
ON
Cohort 1
100%
A1.sarah_admin
2026-04-18
›
discovery.ai_deep_search
Wedge F
Deep-research module on listing
CANARY
Trust+ providers
15%
A1.sarah_admin
2026-05-02
›
commerce.kbank_qr_payment
Wedge F
KBank QR rail (Thai-local)
DEV
Internal only
0%
PR2.backend
2026-05-04
›
community.realtime_ama
Wedge K · sketch
Live AMA module · A4.live moderation
ON
All
100%
A1.sarah_admin
2026-04-12
›
insight.predictive_ml_v2
Wedge H · sketch
Predictive booking demand model
OFF
—
0%
A2.eng_lead
2026-03-28
›
i18n.tolgee_runtime_v2
Wedge F
Tolgee 2.x runtime + AI translation
CANARY
Per-locale
62%
A2.platform
2026-05-08
›
Recent flag changes · audit logFull audit log →
2026-05-08 14:22Z · A1.sarah_admin · i18n.tolgee_runtime_v2 rollout 45% → 62%
2026-05-04 09:14Z · PR2.backend · commerce.kbank_qr_payment created · DEV scope
2026-05-02 11:08Z · A1.sarah_admin · discovery.ai_deep_search rollout 5% → 15%
2026-04-18 16:32Z · A1.sarah_admin · quotation.ai_match_v2 rollout 50% → 100% · GA
📋 SCENARIO A2 platform engineer watches per-tenant unit-economics (LLM tokens, storage, egress, support minutes) against the cost guardrail. Outliers trigger an autoresponse: throttle, alert tenant owner, and open an incident ticket for an A1 commercial review.
83Cost dashboards (F.X6)
/dashboard/ops/cost

Cost · last 30 days

F.X6 lock · multi-tenant cost attribution · forecast · alerting
⚠ LLM cost approaching cap
88% of $9,500 monthly cap used · 12 days remaining in cycle · top driver: Bumrungrad ai_match_v2 traffic ↑ 38% wow
Investigate →
Total spend
$24,840
vs $32k budget · ↑ 6% MoM
AI / LLM
$8,420
⚠ 88% of cap
Compute
$6,840
stable
Vendor (Stripe etc.)
$4,180
↑ with GMV
End-of-month forecast$200 over budget
$24,840 spent12d remaining$32,000 budget
Projected total · end May$32,200 (↑ 3% MoM)
Active alerts
LLM cost > 85%FIRING
DB CPU > 80%FIRING
Stripe error rate > 1%ok

Per-tenant attribution

Compute + AI usage allocated to provider tenants
Tenant
Compute
AI / LLM
Total / 30d
% of total
Bumrungrad
TH · Trust+
$420
$1,420
$1,840
7.4%
›
BNH Hospital
TH · Trust+
$240
$680
$920
3.7%
›
Bangkok Hospital Phuket
TH · Free
$180
$460
$640
2.6%
›
MedPark
TH · Free
$120
$240
$360
1.4%
›
2,443 other tenants
avg $8 / mo
$5,880
$5,620
$21,080
84.9%
›

By service

Service
Compute
DB
AI
Total
ai-service
$420
$120
$8,420
$8,960
›
product-core
$3,840
$2,140
—
$5,980
›
platform-services
$1,420
$680
—
$2,100
›
async-workers
$680
$240
—
$920
›
📋 SCENARIO A2 site reliability engineer monitors SLO attainment and error-budget burn for booking, payment, and chat surfaces. Burn-rate alerts page the on-call rotation; sustained breaches freeze feature rollouts until a postmortem and remediation plan land.
84SLO health dashboard (F.P5)
/dashboard/ops/slo

SLO health

F.P5 lock · all Wedge F SLOs · 30d burn-rate · breach history · on-call routing
3 of 4 within budget 1 burning fast
Quotation submit p95
Target < 200ms · 99.9% availability · owner A2.platform
187ms
WITHIN BUDGET
30 days agotoday · burn-rate 0.08×
Error budget: 92% remaininglast breach: never
Token validation p95
Target < 50ms (cache-hit) · owner A2.platform · F-locked: TOTP-bypass fix per WEAKNESS.md
52ms
4% OVER
30 days agotoday · burn-rate 2.4× · trend ↑
Error budget: 22% remaininginvestigate Zitadel cache TTL
Notification dispatch p95
Target < 2s · ReNovu · owner A2.platform
1.2s
WITHIN BUDGET
30 days agotoday · burn-rate 0.21×
Cold-start TTHW (Time To Hello World)
Target < 240s · new tenant onboarding · owner PR2.backend
218s
WITHIN BUDGET
30 days agotoday · burn-rate 0.12×
Recent breaches · 90 days3 breaches · 0 customer-facing
Date
SLO
Duration
Outcome
2026-04-12
Token validation
38m
soft
2026-03-08
Notification p95
14m
recovered
2026-02-22
Quotation submit
22m
recovered
On-call rotationPagerDuty mock
Primary
Until Mon 09:00
A2.eng_lead
Secondary
Until Mon 09:00
A2.platform_eng
Escalation lead
24/7
A1.sarah_admin
📋 SCENARIO A2 platform engineer browses the service catalog to map ownership, dependencies, and on-call rotation per service. They use it to plan a Wedge migration or to find the right team to ping when a downstream service degrades.
85Service catalog (F.X3)
/dashboard/ops/catalog

Service catalog · 6 services

F.X3 lock · service inventory mirrors arokago-v1-mono · per-service ownership · SLO · deploy log · SBOM · Wedge maturity tag
Service
Repo / role
Owner
Recent deploy
SLO
Wedge
ai-service
Python · LiteLLM
arokago-v1-mono / services/ai · LLM gateway · matching · embeddings
A2.eng_lead
A2.platform_eng (sec)
v1.0.8 · 2h ago
96%
F
›
product-core
Django
services/backend/app · Quotation · Commerce · Discovery · Trust
PR2.backend
A2.platform_eng (sec)
v2.1.0 · 6h ago
99.4%
F
›
platform-services
Python · Zitadel
Identity · MFA · Notifications · ReNovu bridge
A2.platform_eng
A1.sarah_admin (sec)
v1.4.2 · 1d ago
98%
F
›
async-workers
Celery · Redis
Tolgee fan-out · embedding jobs · audit-log archive · DR tasks
A2.platform_eng
v0.9.4 · 3d ago
97%
F
›
frontend-app
Next.js 16 · React 19
Patient + provider + admin web · next-intl · Tolgee runtime
A2.eng_lead
v0.4.1 · 4h ago
99.7%
F
›
chat-service
Socket.io · legacy
P1 ↔ provider chat · slated for Chatwoot replacement
PR2.backend
v0.7.2 · 9d ago
99.5%
F→G
›
ai-service · drill-down preview
Click any row above to open · this is the canonical layout.
Wedge F · production
Deploy log · last 7d
14 deploys · 0 rollbacks
Latest: v1.0.8 · A2.eng_lead
Open PRs
12 · 3 awaiting review
2 blocked on tests
Active alerts
4 · 1 critical
LLM cost > 85%
SBOM
247 deps · 0 critical CVE
Last scan 2h ago
📋 SCENARIO A1 compliance reviewer streams the live audit tail to investigate a privileged-action incident — admin role grants, KYC overrides, payout releases. They export the relevant window for the legal hold workflow when the incident touches financial product or PHI.
86Audit log live tail (F.X2)
/dashboard/ops/audit

Audit log

F.X2 lock · most-cited locked capability · A1 + Compliance officer monitor every privileged action · append-only · cryptographically verified
● TAILING
Last 24h · A1.sarah_admin Privileged only Denied 12 Trust events Fraud flags 3
2026-05-09T14:38:02Z · A1.sarah_admin verification.granted provider=bumrungrad trace_id=4bf92f3577…
2026-05-09T14:37:48Z · P1.sarah_k quotation.created proc=dental-implant trace_id=00f067aa0b…
2026-05-09T14:37:32Z · A4.editor moderation.rejected content=community-thread reason=medical-accuracy trace_id=2b4ce91a3f…
2026-05-09T14:37:18Z · P1.james_l auth.mfa_challenge_passed factor=totp trace_id=8e4a2d6f1c…
2026-05-09T14:37:02Z · A1.sarah_admin tenant.access provider=bumrungrad scope=billing trace_id=9d3f7e2b8a…
2026-05-09T14:36:48Z · PR1.bumrungrad offer.sent quotation=q_abc to=sarah_k trace_id=00f067aa0b…
2026-05-09T14:36:32Z · P1.maria_s consent.granted provider=acibadem scope=quotation trace_id=4f8c2e9b7d…
2026-05-09T14:36:18Z · A1.sarah_admin user.impersonated target=marco@bumrungrad reason=support-ticket-1842 trace_id=7f3e1c8d9b…
2026-05-09T14:36:08Z · P1.unknown_192_168 auth.mfa_challenge_failed factor=totp · 3rd attempt · ⚠ rate-limited trace_id=1a4b8c5e2f…
2026-05-09T14:36:14Z · system audit.integrity_check_passed sequence=4928401 [chain ✓]
✓ Sequence integrity verified
Sequence #4928401 · cryptographic hash chain intact · last anomaly: never · verifier runs every 60s · last run 12s ago
4,928,401 entriesSHA-256 + Merklelast verified 12s ago
Per-actor drill-down
Drill into any actor's privileged action history · audit-logged on filter.
Retention
Hot retention30 days
Archive retention7 years
Compliance basisPDPA + GDPR Art. 30
📋 SCENARIO A2 SRE schedules and runs the quarterly DR drill — region failover, restore-from-backup, and game-day chaos test. Pass/fail outcomes feed the resilience scorecard reported to A1 leadership and the board's risk committee.
87DR drill UI (F.X7)
/dashboard/ops/dr-drill

Disaster recovery drills

F.X7 lock · RTO 4h target · RPO < 5min · quarterly cadence · cross-region failover deferred to Phase 0.β
Last drill 2026-04-18 · PASS
RTO target 4h · achieved 2h 14m ✓ · RPO achieved 3 min · scenario: primary DB failover · 0 customer-visible impact
healthy
Last drill
2026-04-18
RTO 2h14m · target 4h ✓
Next scheduled
2026-07-18
Quarterly cadence · scenario TBD
Backup integrity
100%
All snapshots verified · 2h ago
📌 Phase 0.β · Cross-region failover
Single-region today (Singapore primary). Phase 0.β adds Bangkok hot-standby + cross-region replication; this surface will gain region-by-region drill controls.

Drill history

Date
Scenario
RTO actual
RPO actual
Outcome
Lead
2026-04-18
Primary DB failover · simulated
2h 14m
3min
PASS
A2.platform_eng
›
2026-01-22
Region failover · simulated
3h 52m
2min
PASS
A2.platform_eng
›
2025-10-12
Audit log restore from archive
22m
1min
PASS
A1.sarah_admin
›
2025-07-08
Object storage restore (provider images)
1h 04m
5min
SOFT
A2.platform_eng
›
📋 SCENARIO A1 admin manages their own MFA enrolment, IP allow-list, and session policy on the Zitadel-backed account. Changes to MFA factors require step-up authentication and are written to the audit log immediately.
23Admin settings (security & MFA)
/dashboard/settings

Platform settings

Identity primitive · mfauth + Zitadel sync · F-locked: TOTP-bypass fix per WEAKNESS.md commons/auth/totp_auth.py:25
Last login: 2 hours ago · Bangkok, Thailand
Chrome 124 on macOS · IP 110.171.34.18 · matched 14 prior logins from this device
If this wasn't you →

Security & MFA

Hardware key (WebAuthn)
YubiKey 5 NFC · last used 2 hours ago · primary factor
ENABLED
TOTP authenticator app
1Password · last used 2 hours ago · F-locked: bypass fix at totp_auth.py:25
ENABLED
Recovery codes
8 of 10 unused · stored in 1Password vault · 2 used in last 90d

Password policy

Minimum length16 characters
MFA required for admin rolesenforced
Auto-rotation90 days
Reuse windowlast 12 passwords
Breach check (HIBP)enabled

Active sessions

3 devices currently signed in · revoke any device to terminate its session
Device
Last activity
IP / Location
MacBook Pro · Chrome 124
this session
just now
110.171.34.18 · Bangkok TH
current
—
iPhone 15 · Safari
arokago iOS app
3h ago
110.171.34.18 · Bangkok TH
Revoke
Chromebook · Chrome 124
conference room
2d ago
203.150.215.12 · Bangkok TH
Revoke

API keys

Machine credentials · scoped permissions · auto-expire · audit-logged on every request
Name
Scope
Last used
Expires
Analytics ETL
ak_anal_4f8c…
read:analytics · read:audit-log
14m ago
2026-08-04
active
›
DR drill bot
ak_dr_b8e2…
admin:dr-drill · read:slo
2d ago
2026-07-18
active
›
Coordinator portal · staging
ak_pr3_2a91…
read:quotation · write:offer (staging only)
19d ago
2026-05-14
5d
›
📋 SCENARIO Admin reviewer (A1) configures the trust-signal stack — JCI accreditation, medical-board verification, GBP rating sync, ArokaStars, and verified-booking review weighting. Weight changes require an A1 + A2 paired sign-off before flipping live.
30Trust-signal stack admin (A1)
/dashboard/trust/verification

Trust-signal stack admin

A1 Platform Operator · V3 verification scheme · 1,247 providers verified · 23 expiring in next 90d · 3 revocations pending appeal · spec L29 + L82 + L155
23 expiring 90d 3 appeals
Status Active 1,247 Expiring 23 Expired 8 Revoked 3 Pending 47 Trigger License expiry 2yr cycle Manual
📌 Phase 2+ · Tiered + scoped verification
Bronze / silver / gold tiers + per-procedure scope deferred. v1 binary verified shield active across all listings · spec L57-61.
Provider
Archetype
Granted
Expires
Trigger
A1 owner
Re-verify q
Bumrungrad International
JCI · ISO 9001
Hospital group
2024-08-12
2026-08-12
2yr cycle
A1.sarah_admin
#3
›
Bangkok Dental Hospital
Dental · MCT registered
Specialty clinic
2024-11-04
2026-07-04
License expiry
A1.marco_g
#1
›
Phyathai Sriracha
Hospital · JCI
Hospital group
2025-02-19
2027-02-19
2yr cycle
A1.sarah_admin
—
›
BNH Hospital
JCI · TH-MoPH
Hospital group
2024-06-30
2026-06-30
2yr cycle
A1.marco_g
#7
›
Samitivej Sukhumvit
JCI · pediatric center
Hospital group
2025-04-02
2027-04-02
2yr cycle
A1.sarah_admin
—
›
Vejthani Hospital
⚠ Appeal pending
Hospital group
2024-12-15
REVOKED
Manual
A1.sarah_admin
—
›
Khon Kaen Ram Hospital
JCI · TH-MoPH
Hospital group
2025-01-22
2027-01-22
2yr cycle
A1.marco_g
—
›
BNH Hospital Bangkok
JCI · cardiology focus
Hospital group
2024-09-08
2026-09-08
2yr cycle
A1.sarah_admin
#5
›
Bumrungrad International · drillVERIFIED · expires 2026-08-12
Hospital group · 2yr cycle re-verify · queue position #3
Audit log · grant / revoke / re-verify
2024-08-12 · verification.granted · A1.sarah_admin · evidence: JCI 2024 + ISO 9001
2024-11-30 · verification.evidence_added · A1.sarah_admin · MCT registration #BMR-1247
2025-04-18 · verification.spot_check_passed · A1.sarah_admin · annual
2026-05-09 · verification.expiry_alert · system · 90d before expiry
Third-party feed augment
JCI accreditation listing✓ active
ISO 9001 database✓ active
Thai MCT registration✓ active
A1 review notes

2025-04 spot-check passed without findings. Hospital expanding cardiac wing · re-verify due 2026-08 may include scoped procedure expansion review.

Bulk re-verification · 23 providers expiring
Distribute workload across A1 team (sarah_admin · marco_g · 2 more) · estimated 6.5 days at 4 reviews / day / reviewer.
⛨ All actions logged immutably
Every grant / revoke / re-verify writes to append-only audit log · cryptographic verification (Merkle tree · SHA-256) · compliance.md L34 · view audit-log live tail →
📋 SCENARIO A4 moderator reviews provider-uploaded before/after photos against signed patient consent records. Missing or expired consent triggers an automatic take-down and a provider warning; repeat offences escalate to A1 for tier downgrade.
31A4 photo-consent review queue
/dashboard/moderation/photo-consent

A4 photo-consent queue

A4 Moderator · review + editorial sub-tiers · gates photo gallery uploads + manipulation detection · spec L34 + L60 + L110 acceptance #5-6
47 pending 12 AI-flagged 3 expired-consent
Pending review (47) AI-flagged (12) Approved today (84) Rejected today (3) Expired-consent requests (3)
24 items · clear-AI · low-risk
All have ✓ patient consent + clean AI manipulation check + jurisdiction-allowed.
photo
thumb
Bumrungrad · before/after gallery✓ patient consented
Context: gallery upload · uploader: PR2 dr.anil@bumrungrad · dental implants · 4 photos
AI: ✓ clean jurisdiction: 🇹🇭 TH allowed
⚠ AI
flagged
BNH Hospital · review BK-A8C2 photo⚠ pending consent · awaiting P1 response
Context: review BK-A8C2 photo · uploader: P1 maria_s · cosmetic before/after
AI: 84% beauty-filter detected 12% deepfake possible jurisdiction: 🇰🇷 KR-blocked · cosmetic before/after
photo
thumb
Bangkok Dental · gallery photo + caption✓ patient consented
Context: gallery upload · uploader: PR2 dr.somsak@bkkdental · single tooth implant
AI: ✓ clean Caption in TH · A4.translation review pending
⛨ Cross-language moderation · auto-translated by Tolgee · A4.translation reviews per locale before publish · spec trust-and-safety.md L44
12 AI-flagged photos · click any to open fullscreen review panel with manipulation overlay.
BNH Hospital · cosmetic before/after84% beauty-filter
Open fullscreen → original photo · AI overlay highlighting filter regions · consent text · jurisdiction rules table
Phuket Plastic · review photo92% deepfake
Reverse-image search returns no source · escalate to A1 recommended.
84 approved today · A4.review · all within 24h SLA · audit-logged.
3 rejected today · all uploaders notified with revision notes · 7d resubmit window.
3 photos served past consent expiry · re-issue requests sent to patients · gallery temporarily hidden.
Vejthani · 2024 dental galleryconsent expired 2026-04-30
3 photos · re-consent email sent to P1 sarah_k · 7d response window · gallery hidden until consent renewed.
Provider-flagged review escalations
Per pr03 footer · provider defamation appeals route to A4.review queue.
⛨ All decisions logged immutably
Approve / reject / hold / escalate writes to append-only audit log · response SLA 24h per spec G L110 acceptance #6 · cross-link provider gallery upload flow →
📋 SCENARIO A1 and A2 jointly monitor Google Business Profile OAuth health, review-import lag, and rating-drift anomalies across providers. Stalled OAuth tokens are surfaced for re-consent; suspicious review surges are flagged to A4 moderation.
32GBP sync admin (A1 + A2)
/dashboard/integrations/gbp

GBP sync admin

A1 + A2 administer Google Business Profile sync infrastructure · 847 providers connected · 412k Google reviews indexed · 12 sync failures in last 24h · API quota: 87% used today · spec L32 capability 5 + L138 risk
847 connected 12 failures quota 87%
Providers connected
847
+12 last 7d
Reviews synced
412k
+1.4k last 24h
Last full reconcile
2h ago
2026-05-09 12:38 UTC
API quota
87%
OAuth-grant pending
14
awaiting provider auth
📌 Sync direction · read-only at G
V3 ← Google · ingest reviews into ArokaGO trust panel. Write-back (V3 → Google) deferred to Phase 2 · review-duplication risk mitigation per spec L120 (open question resolved).
read-only · v1
Provider
OAuth
Last sync
Reviews
Errors
Bumrungrad International
place_id ChIJ-bmrgrd
granted
2h ago
14,287
—
›
Bangkok Dental Hospital
place_id ChIJ-bkkdent
granted
2h ago
3,142
—
›
BNH Hospital
place_id ChIJ-bnh-bkk
expired
14h ago
8,491
OAuth
›
Vejthani Hospital
place_id ChIJ-vejthn
revoked
2d ago
2,184
revoked
›
Phyathai Sriracha
place_id ChIJ-phytsr
granted
2h ago
5,672
—
›
Khon Kaen Ram Hospital
place_id ChIJ-kkr-th
granted
5h ago
11,824
rate-lim
›
Samitivej Sukhumvit
place_id ChIJ-samtsv
granted
2h ago
9,358
—
›
BNH Hospital Bangkok
place_id ChIJ-bnhhsp
granted
2h ago
4,201
—
›
Phuket International
place_id ChIJ-pktintl
expired
9h ago
3,847
OAuth
›
Yanhee Hospital
place_id ChIJ-yanhee
granted
2h ago
2,914
—
›
Mount Elizabeth Novena
place_id ChIJ-mtelizn
granted
2h ago
7,628
—
›
Gleneagles Bangkok
place_id ChIJ-glnbkk
granted
2h ago
5,047
—
›
BNH Hospital · drillOAuth expired 14h ago
Last successful sync: 2026-05-08 22:14 UTC · 8,491 reviews cached
Sync log
2026-05-09 02:18 · gbp.sync_failed · token_expired · retry_after=oauth_regrant
2026-05-08 22:14 · gbp.sync_ok · 47 new reviews · 412 ratings updated
2026-05-08 16:14 · gbp.sync_ok · 12 new reviews
2026-05-08 10:14 · gbp.sync_ok · 8 new reviews
OAuth scopes granted
business.manageexpired
reviews.readexpired
Sync failures · last 24h · 12 total
OAuth token expired
8
BNH Hospital · Phuket Intl · 6 more
Provider revoked permission
2
Vejthani · 1 more
Google API rate-limit
2
Khon Kaen Ram · 1 more · auto-retry 1h
⛨ Graceful degradation · 24h cache TTL
Cached Google reviews continue serving on sync failure · alert sent to PR1 if sync down >48h · spec L138 mitigation.
📋 SCENARIO A4.review moderator screens patient reviews tied to a verified booking — checking for PHI leakage, retaliatory tone, or paid-review patterns. Approved reviews publish to the provider profile; flagged ones trigger a provider right-to-respond workflow.
34A4.review moderation queue (verified-booking reviews)
/dashboard/moderation/reviews

A4 review moderation queue

A4.review sub-tier · moderates verified-booking reviews + handles defamation escalations from PR1/PR2/PR3 · spec L60 + trust-and-safety.md L41 + L42
47 pending 12 appealed 3 fake-flagged
Pending light-review (47) Provider-appealed (12) AI fake-pattern flagged (3) Helpful-vote anomaly (5) Approved today Rejected today
★★★★★ "Excellent dental work · staff was kind"+92% positive sentiment
P1 sarah_k · Bumrungrad · Dental implants single tooth · slot 2026-04-18 · BK-A8C2

"Excellent dental work and the staff was incredibly kind. Dr. Anil walked me through every step. Would recommend to anyone considering implants in Bangkok."

verified booking ✓ P1 author · 2 prior reviews · all positive
★★★★☆ "การรักษาดีมาก แต่รอนาน"Review in TH
P1 niran_t · Bangkok Dental · Root canal · slot 2026-04-22 · BK-B3D9

"การรักษาดีมาก แต่รอนาน 45 นาที สำหรับนัดที่จองไว้แล้ว · medical-content edge case: mentions wait-time delay"

verified booking ✓ A4.translation reviewed for medical-content · spec G L88
⛨ Per-locale moderation indicator · A4.translation reviewed for medical-content edge cases (wait-time framing) · trust-and-safety.md L44
★☆☆☆☆ "Worst experience · doctor was rude · botched job"−83% likely defamation
P1 anon-userX · BNH Hospital · Rhinoplasty · slot 2026-04-12 · BK-C7F1 · provider-appealed 2026-05-04

"Worst experience of my life. Dr. Somchai was rude and the result is a botched job. They scammed me out of $4,800. AVOID THIS PLACE!!!"

Author history
P1 anon-userX · first review · account 4d old · IP signals match prior bot-cluster (12% confidence) · strike record · 1 prior strike (2026-03 community spam) · escalation level: medium
Provider response (visible)

"We're sorry to hear about your experience. Our records show successful follow-up consultations on April 14 and April 21. Please reach out to our patient relations team."

Provider's defamation-appeal evidence
Pre-op consent signed✓ on file
Post-op review at day 14✓ patient signed off · "satisfied"
Refund / dispute historynone filed
Photo evidence · result match✓ within agreed outcome envelope
3 reviews matched AI fake-review-pattern detector · cluster of 5-star reviews from accounts <48h old.
"Phyathai · ★★★★★ amazing service"94% fake-pattern
3 of 5 cluster matches · same phrasing template · accounts 24-48h old · escalate to A1 recommended.
5 reviews show abnormal helpful-vote velocity (sudden spike from low-trust accounts) · vote-fraud heuristics triggered.
142 approved today · all within 24h SLA · audit-logged.
8 rejected today · all authors notified with revision notes · 7d resubmit window.
Defamation legal escalation queue
Escalations route to A1 + legal · structured evidence packet auto-assembled.
⛨ Cross-platform reputation portability
Strike record persists across ArokaGO surfaces (community + reviews) · escalation tier raises with each strike · trust-and-safety.md L46 · cross-link provider review-flag flow →
📋 SCENARIO A1 ops manager reviews aggregated capacity utilisation — chair-hours, OR slots, doctor availability — across all PR5 multi-clinic groups. They identify under-booked slots and notify ops leads to push promotions or open them to ad-spend boosts.
35PR5 Ops Manager · capacity rollups (PR5 first appearance)
/dashboard/ops/capacity

Operations dashboard

PR5 Ops Manager · Bumrungrad Group · 4 branches · 47 practitioners · 1,247 slots/week · 87% utilization · spec L62 (PR5 introduced at G)
87% utilization 1 capacity alert
⛨ PR5 Ops Manager · multi-branch capacity planning + slot inventory oversight
RBAC scope: read across all 4 branches · write capacity-plan + slot-inventory · cannot edit clinical records or pricing. First introduced at Wedge G per spec L62.

Branch rollup · last 30d

Branch
Slots offered
Booked
Util %
Cancel %
Revenue
Top procedure
Bumrungrad Sukhumvit
flagship · 24 practitioners
3,840
3,341
87%
4%
$1.2M
Cardiac surgery
›
Bumrungrad Phyathai (Dental)
14 practitioners · ⚠ saturated
2,016
1,976
98%
2%
$612k
Dental implants
›
Bumrungrad Phuket
resort branch · 6 practitioners
864
691
80%
8%
$284k
Cosmetic
›
Bumrungrad Chiangmai
satellite · 3 practitioners
432
298
69%
5%
$94k
Wellness checkup
›
🚨 Bumrungrad Phyathai dental · 98% utilization next 14d
Demand exceeds available slots. Suggested action: +3 slots/day Mondays-Wednesdays · estimated revenue uplift $84k / month.

Slot inventory model

Spec L119 · open question resolved at G as per-staff (Cal.com native). Per-treatment-room (capacity-based) deferred to Phase 2.
Per-staff (Cal.com native)v1 default
Each practitioner manages own calendar. Cal.com primitive. Active across all 4 branches.
Per-treatment-room (capacity-based)Phase 2
Forward-compat for surgical theaters · constrained by physical room availability rather than practitioner availability.

30d demand forecast

📈 Demand forecast: 47% above current capacity
Driven by Bumrungrad Phyathai dental + Sukhumvit cardiac. Projected 1,832 incoming requests vs 1,247 weekly slots.
✓ 0 multi-branch slot conflicts detected
All 4 branches calendar-synced cleanly · no double-booked practitioners across Sukhumvit / Phyathai / Phuket / Chiangmai · last sync 2m ago.

Refund-window cohort · last 30d

Ties to spec L37 refund-window logic · monitor cancellation patterns relative to commerce refund tiers.
Cancelled · full-refund window
4%
low-friction · expected
Cancelled · 50% window
7%
+2pp vs prior 30d
Cancelled · no-refund window
1%
friction working
Refund-revenue impact
$2.4k
last 30d · -$840 vs prior
Per-practitioner slot configuration
Drill into individual practitioner Cal.com calendars across branches · PR2 self-serves; PR5 has read + bulk-edit RBAC.
📋 SCENARIO A2 ML engineer trains and evaluates the v3.2 lead-scoring model — feature drift, AUC, calibration by archetype, fairness by locale. A successful run gets promoted to a shadow deploy; bias-audit failures block promotion until fixed.
40Lead-scoring model · v3.2 training UI
/admin/models/lead-scoring

Lead-scoring model · v3.2

A2 Platform Analyst · last trained 2026-05-08 02:00 UTC · production · spec H L121-122 (nightly run, dormant reactivation) · V2.5 predictive_models skeleton wired in
production · v3.2 nightly 02:00 UTC
⛨ A2 Platform Analyst · model lifecycle owner · gradient-boosted classifier
RBAC: train + promote + rollback · cannot view PII attributes (anonymised feature vectors only) · drift alarms route to A2 on-call · provider-side score breakdown surfaced read-only in PR3 lead detail.

Model card · v3.2 production

Algorithm
GBM · XGBoost
gradient-boosted classifier
Features
24
7 engagement · 6 profile · 5 budget · 6 context
Training set
18,400
leads · 90d window · stratified
Holdout AUC
0.83
+0.04 vs v3.1
Drift score (PSI)
0.08
stable · <0.10 threshold
Last deployed
2026-05-08
A2 promoted · 02:14 UTC
Calibration
0.94
Brier score 0.11 · isotonic

Feature importance · top 15

SHAP-derived weights · normalised to sum 1.0 across all 24 features.
Engagement: response time within 1h
0.18
Profile match score (procedure × provider tier)
0.14
Budget signal (declared range vs procedure med.)
0.12
Travel-window proximity (≤30d)
0.10
Search→inquiry depth (pages / sessions)
0.08
Quotation acceptance (P1 historical)
0.06
Cross-border verification (passport · medical visa)
0.05
Insurance pre-auth status
0.05
Locale match (P1 ↔ provider language)
0.04
Returning patient flag
0.04
Channel preference signal (WhatsApp/LINE/email)
0.03
Referral source quality (organic vs paid)
0.03
Geo-distance to nearest branch
0.02
Account age (days since signup)
0.02
Device class (mobile vs desktop)
0.02

Version history · promote · rollback · A/B compare

Version
AUC
Brier
Drift
Status
Trained
v3.2
+travel-window feature
0.83
0.11
0.08
production
2026-05-08
v3.1
+returning-patient flag
0.79
0.14
0.06
staged
2026-04-12
v3.0
initial V3 baseline
0.74
0.18
0.12
archived
2026-02-04
Per-archetype weight overrides6 archetypes
Wellness providers de-weight travel-window (resort + extended-stay context); Healthcare emphasises insurance pre-auth.
Wellness · travel-window weight0.10 → 0.04
Healthcare · insurance pre-auth0.05 → 0.09
Diagnostic · profile-match weight0.14 → 0.18
Pharmacy · returning-patient flag0.04 → 0.11
⚠ Risk panel · false-positive trust impact23 reverses
Per H risk L144: false-positive scoring degrades coordinator trust. Below = PR3 manual reverses (provider-side score override) in last 7d.
PR3 reverses last 7d23
Avg score on reversed leads0.81 (high)
Top reverse reason"obvious tire-kicker"
Suggested feature auditengagement signal · 1h-response
Provider-side score breakdown
PR3 sees per-lead score + top 3 contributing features in inbox + lead detail. Read-only with reverse action.
📋 SCENARIO A2 data scientist tunes per-archetype segmentation models (cosmetic, dental, IVF, wellness, etc.), reviews cluster stability over time, and ships updated definitions to the marketing automation pipeline. Stale clusters trigger a retrain queue notification.
41Segmentation models · per archetype
/admin/models/segmentation

Segmentation models · 6 archetypes · 18 active · 4 in training

A2 owns per-archetype segmentation tuning · providers consume promoted models via PR3 segments library (pr07).
18 active 4 in training
⛨ Insight H · per-archetype tuning · prevents one-size segmentation
Healthcare patients differ from wellness patients differ from pharmacy patients. A2 tunes feature weights per archetype, then promotes to provider segment library where PR3 selects from curated set.
Healthcare (5 models) Wellness (4) Diagnostic (3) Insurance (2) Facilitator (2) Pharmacy (2)
Segment model
Members
Drift (PSI)
Last trained
Status
High-intent inquirer
budget+travel-window+depth signals
12,400
0.07
2026-05-08
promoted
›
Returning patient
prior-booking + loyalty + cadence
3,840
0.05
2026-05-06
promoted
›
Price-sensitive
budget-elasticity + comparison-shopping
8,200
0.11
2026-04-28
drift alarm
›
Premium-tier patient
LTV + procedure-tier + concierge signals
1,420
0.04
2026-05-02
promoted
›
Cross-border traveler
medical-visa + multi-leg trip + locale
5,600
training
2026-05-09
in training
›
Segment model
Members
Drift (PSI)
Last trained
Status
Resort retreat seeker
extended-stay + leisure-overlay
4,120
0.06
2026-05-04
promoted
›
Burnout recovery
stress markers + leave-of-absence signals
2,640
0.08
2026-05-01
promoted
›
Longevity tracker enthusiast
wearable-data + biomarker testing
1,180
0.09
2026-04-30
promoted
›
Group retreat (2-6 ppl)
multi-pax bookings + couple/family overlay
980
training
2026-05-09
in training
›
3 models: Annual-checkup recurrer · Specialty-imaging hunter · Genetic-screening curious. Click any to open feature breakdown + drift heatmap. Drift across all 3 stable (PSI < 0.10).
2 models: Pre-auth-required cohort · Out-of-network shopper. Both promoted to provider library. Insurance archetype is read-heavy on quotation history.
2 models: Repeat-relay coordinator · One-off-introducer. Used for facilitator referral commission tiering.
2 models: Chronic-medication refiller · Travel-medication tourist. Pharmacy archetype is the smallest active provider pool (V3.2).

Cluster visualisation · UMAP 2D · Healthcare archetype

Coloured by predicted segment · click any cluster to inspect representative members + feature centroid.
High-intent
(12.4k)
Returning
(3.8k)
Price-sensitive
(8.2k) ⚠
Premium
(1.4k)
Cross-border
(5.6k)
UMAP n_neighbors=15 · min_dist=0.10 · 2D projection · sampled 5%

Promote model → provider segment library

Promote: Cross-border traveler model → all Healthcare providers
Once promoted, segment appears in PR3 segments library across all 84 active healthcare providers · provider sees model name + member-count preview · cannot edit weights but can subscribe/dismiss per workspace.
Provider segment library (PR3)
PR3 selects from promoted segments + saves cohorts for drip campaigns. Read-only on weights · subscribe/dismiss per workspace.
📋 SCENARIO A4.editorial moderator reviews provider marketing copy, ad creatives, and email/LINE blasts before they enter the queue. They check for medical-claim compliance, regulated-keyword usage, and country-specific disclaimers; rejected drafts return with rubric notes.
42A4.editorial · campaign-copy review queue
/admin/editorial/queue

Editorial review · 12 pending · SLA 24h · 4 nearing breach

A4.editorial sub-tier · spec H L70 · gates campaign copy when provider-authored content references medical claims · risk: AI hallucination liability (H L145).
4 nearing SLA breach 8 within SLA
⛨ A4.editorial · campaign-copy gate · medical-claims aware
RBAC: review + edit-and-return + approve-with-edits · cannot publish without review · senior tier curates banned-phrase catalog. Routes here when AI-generated or when copy contains medical-claim phrase patterns.

Pending review queue (12)

Submission
Provider
Campaign
Channel
Authored
SLA countdown
Assignee
SUB-7841
2026-05-09 14:22
Bumrungrad
PR3 · marketing
Knee replacement · post-op recovery
drip · 5 emails · 4 locales
Email
⚠ AI
2h 14m
Sarah K. (A4)
›
SUB-7842
2026-05-09 11:47
BNH Hospital
PR3 · ads
Cosmetic · safe-recovery messaging
3 ad variants · EN+TH
LinkedIn ad
Human
8h 02m
Aksu Y. (A4)
›
SUB-7843
2026-05-09 09:12
Bangkok Dental
PR3 · marketing
Implant aftercare · 30d follow-up
WhatsApp template · TH
WhatsApp
⚠ AI
10h 47m
Niran T. (A4)
›
SUB-7844
2026-05-08 22:30
Samitivej
PR3 · marketing
Pediatric vaccination reminder
LINE template · TH+EN+JA
LINE
Human
21h 14m
unassigned
›
SUB-7845
2026-05-08 19:08
Phyathai 2
PR3 · marketing
IVF · success-rate landing copy
email + landing · EN+TH+ZH
Email
⚠ AI
14h 22m
Sarah K. (A4)
›
Review · SUB-7841 · Bumrungrad · Knee replacement post-op recovery
5-email drip · 4 locales (EN/TH/AR/JA) · AI-authored by PR3 marketer · submitted 2026-05-09 14:22 · SLA breach in 2h 14m
EN TH AR JA ZH KO VI ID FR DE ES TR
Email 1 of 5 · subject + body · auto-flag overlay enabled.
Subject: Your knee will feel new again — a 30-day recovery roadmap from Bumrungrad.
Dear [first_name],
Your total knee replacement was a success. Now we'll guide you to a guaranteed full recovery — Bumrungrad's recovery program has helped 98% of patients return to normal activity within 30 days.
Day 1-7: Gentle range-of-motion exercises. Day 8-14: walking with support. Day 15-30: a complete cure begins.
— Auto-disclosed: this email contains AI-assisted content per H spec L96.
2 banned phrases (red) 1 medical-claim flag (amber) AI-disclosed
TH copy · Tolgee auto-translation reviewed · 1 banned phrase flagged · click EN tab for full review pattern.
AR copy · RTL preview · 0 banned phrases · 1 disclosure missing.
JA copy · 0 banned phrases · honorific check passed.
ZH locale not yet translated by provider · Tolgee draft auto-generated for review.
KO locale not yet requested.
VI locale not yet requested.
ID locale not yet requested.
FR locale not yet requested.
DE locale not yet requested.
ES locale not yet requested.
TR locale not yet requested.
Reviewer decision · Sarah K. · A4.editorial
Banned-phrase catalogA4 senior tier only
Per-jurisdiction catalog · enforced per H L96 (US/EU/TH/JP/SA medical advertising rules).
"guarantee" + variantsUS · TH · SA
"cure" (claims absolute)all 12
"miracle" + variantsall 12
"99% / 100% success" w/o sourcerequires citation
"FDA approved" false-claimUS · auto-block
"approved by Ministry"TH · cite source
Reviewer throughput · last 7dSLA met 92%
Sarah K.47 reviewed · 4.2h avg · 96% SLA
Aksu Y.38 reviewed · 5.8h avg · 91% SLA
Niran T.32 reviewed · 3.1h avg · 100% SLA
Approval rate68% (33% w/edits)
Provider CSAT (post-edit)4.4 / 5
Provider-side review queue (PR3)
PR3 sees their submissions + diffs from A4.editorial · accept-edits or counter-edit (returns to A4 for re-review).
📋 SCENARIO A2 experimenter launches and monitors A/B tests across surfaces (search ranking, ad slots, email subject lines). They watch significance, sample-ratio mismatch, and guardrail metrics; failing experiments are auto-paused once SRM crosses threshold.
43A/B framework · consolidated
/admin/experiments

A/B framework · 47 active experiments · 18 platform · 29 provider-scoped

A2 owns A/B engine per H acceptance L118 · V2.5's two coexisting systems (OfferABTest/OfferVariant + platform ABTest*) consolidated to single framework · legacy data migrated 100%.
migration · 100%
✓ Legacy OfferABTest/OfferVariant + platform ABTest* consolidated
12,400 historical assignments preserved · 47 in-flight experiments rebound · audit log at /admin/migrations/ab · A2 sign-off 2026-04-22.
Platform-scoped (18) Provider-scoped (29) Cohort-scoped (8) Scoping policy Audit log
Name
Scope
Sample
Status
Winner
Started
Stat-sig
Inquiry intake form · 3-step vs 1-step
conversion to first-message
platform
12,400
running
3-step (+8%)
2026-04-12
p<0.01
›
Search ranking · v2.4 vs v2.5
listing CTR + first-message rate
platform
8,200
running
v2.5 (+12%)
2026-04-28
p=0.06
›
Trust badge layout · A vs B
listing → inquiry conversion
platform
5,400
running
B (+3%)
2026-05-02
underpowered
›
Quotation reply CTA · A/B/C
P1 reply rate
platform
14,800
stopped
C winner
2026-03-08
p<0.01
›
29 provider-scoped experiments · self-served by PR3/PR5 · subject to scoping policy below.
Name
Provider
Sample
Status
Winner
Started
Email subject · A vs B
Bumrungrad
2,400
running
A (+11%)
2026-05-04
›
WhatsApp opener · A/B/C
BNH Hospital
1,800
running
B (+6%)
2026-05-02
›
Landing CTA color
Bangkok Dental
980
underpowered
—
2026-05-08
›
Drip cadence · 7d vs 14d
Samitivej
3,200
promoted
7d (+18%)
2026-04-12
›
8 cohort-scoped experiments · target specific archetypes (e.g., Healthcare-only, Wellness-only) · A2 + provider co-owned. Inspect to see assignment rules.
Platform-vs-provider scoping policy
A2 sets which experiments require platform approval vs are provider-self-served.
Search ranking changesplatform-only · A2 lead approval
Inquiry intake flowplatform-only · A2 + design
Trust badge / patient-facingplatform · A2 · design
Provider listing copy / photosprovider-self-served
Provider drip cadence + contentprovider-self-served · A4.editorial gate
Provider WhatsApp/LINE openerprovider-self-served · A4.editorial gate
Pricing experimentsplatform-only · CFO + A1 approval
Timestamp
Event
Actor
Experiment
Outcome
2026-05-09 14:22
Promote · 7d drip cadence
A2.sarah_k
Samitivej drip A/B
success
2026-05-08 09:14
Stop · sequential test triggered
A2.system
Quotation CTA A/B/C
winner C
2026-05-04 16:48
Rollback · trust badge B
A2.alex_p
Trust badge A/B
rollback
2026-05-02 11:12
Start · 5,400 sample · 14d window
A2.alex_p
Trust badge A/B
started
2026-04-22 03:14
Migration finalised · 12,400 records
A2.system
— (consolidation)
100%
Inquiry intake form · 3-step vs 1-step
Platform-scoped · 12,400 sample · running · started 2026-04-12 · A2 owner Sarah K.
Variant A · 1-step
14.2%
first-message rate · 6,200 users
Variant B · 3-step ★ winner
22.8%
+8.6pp · 6,200 users · p<0.01
Power · sequential test
99%
stop signal triggered · ready to promote
Per-archetype breakdown
Healthcare lift = +9.2pp · Wellness +12.4pp · Diagnostic +4.1pp · Insurance +6.8pp · Facilitator no diff · Pharmacy +2.0pp.
Provider-side experiments dashboard
PR3/PR5 see their own provider-scoped experiments + cohort breakdowns + history.
📋 SCENARIO A1 compliance officer maintains the per-channel × per-country gating matrix — what marketing surfaces (email, LINE, WhatsApp, ads) require which disclosures by jurisdiction. Updates push to the campaign launcher and block non-compliant sends at submission time.
44Compliance gates · per-channel × per-country
/admin/compliance/gates

Compliance · per-channel + per-country · 4 channels × 12 countries = 48 rule sets

A2 + A4 + legal counsel co-configure · PR2 sees overlay automatically applied at send time · spec H L96 + risk L142 · last reviewed 2026-04-22 by [legal counsel · K. Hanley].
48 rule sets · live last review 2026-04-22
⛨ Compliance H · per-channel + per-country rule engine · Trust & Safety ownership
RBAC: A2 + A4 senior + legal counsel edit · PR2/PR3 read-only with auto-block warnings at send time · all edits two-person sign-off · audit-logged.

Country × channel matrix · click any cell to open rule detail

Green = open · Amber = rules apply · Red = channel restricted. 12 countries × 4 channels = 48 cells.
Country
Email
WhatsApp
LINE
V3 chat
🇺🇸 United States
⚠ CAN-SPAM
✗ TCPA opt-in
— n/a
✓ open
🇪🇺 European Union
⚠ GDPR
⚠ GDPR opt-in
— n/a
⚠ GDPR
🇬🇧 United Kingdom
⚠ PECR
⚠ ICO opt-in
— n/a
✓ open
🇹🇭 Thailand
⚠ PDPA
✓ opt-out ok
✓ open
✓ open
🇯🇵 Japan
⚠ APPI
— low usage
✓ primary
⚠ APPI
🇸🇦 Saudi Arabia
✗ A4 review
⚠ MoH rules
— n/a
⚠ MoH rules
🇮🇩 Indonesia
⚠ UU PDP
✓ open
— n/a
✓ open
🇮🇳 India
⚠ DPDP 2023
⚠ TRAI DLT
— n/a
✓ open
🇦🇺 Australia
⚠ Spam Act
⚠ Privacy Act
— n/a
✓ open
🇨🇦 Canada
⚠ CASL opt-in
⚠ CASL opt-in
— n/a
✓ open
🇸🇬 Singapore
⚠ PDPA
✓ open
— n/a
✓ open
🇰🇷 South Korea
⚠ PIPA
— low usage
⚠ KCC rules
⚠ PIPA
🇺🇸 United States · Email · CAN-SPAM Act
Applicable: CAN-SPAM 2003 + state laws (CA, FL). Banned: deceptive headers, misleading subject lines, no-reply addresses without unsubscribe. Required disclosures: physical postal address + clear unsubscribe link in every commercial email. Default: opt-out (not opt-in). Enforcement: warn-only at send + auto-block if unsubscribe missing.
🇺🇸 United States · WhatsApp · TCPA opt-in
Applicable: TCPA + FCC. Required: explicit prior express written consent for marketing messages. Default: opt-in only. Enforcement: auto-block if no opt-in record · class-action exposure $500-$1500 per message. PR2 cannot send without verified consent record.
🇺🇸 United States · V3 chat · open
First-party conversation initiated by patient · no extra opt-in beyond V3 ToS. HIPAA applies if PHI exchanged · A4.editorial gates copy with medical claims.
🇪🇺 European Union · Email · GDPR + ePrivacy
Applicable: GDPR Art. 6 + ePrivacy Directive. Required: lawful basis (typically consent) · clear unsubscribe · data-controller identity. Default: opt-in. Enforcement: warn-only at send + audit-trail required · DPO sign-off for new templates.
🇪🇺 European Union · WhatsApp · GDPR + ePrivacy
Required: opt-in + double opt-in for medical context · DSAR right + 24h response SLA · enforcement: auto-block without consent record.
🇪🇺 European Union · V3 chat · GDPR scoped
Patient-initiated · Art. 6(1)(b) contract performance basis · 30d retention default unless escalated to Stars review.
🇬🇧 United Kingdom · Email · PECR + UK GDPR
Applicable: PECR + UK GDPR. ICO enforcement. Required: opt-in + clear unsubscribe + identity. Soft opt-in allowed for existing-customer re-engagement only.
🇬🇧 United Kingdom · WhatsApp · ICO opt-in
Same as PECR · ICO enforces · auto-block without consent.
🇬🇧 United Kingdom · V3 chat · open
Patient-initiated · UK GDPR contract basis · 30d retention default.
🇹🇭 Thailand · Email · PDPA
PDPA 2019. Lawful basis: consent or contract. Required: TH-language disclosure + clear unsubscribe. Default: opt-in for marketing · opt-out ok for transactional. Enforcement: warn-only.
🇹🇭 Thailand · WhatsApp · open with PDPA overlay
PDPA applies but channel itself permitted · opt-out allowed · TH-language preferred.
🇹🇭 Thailand · LINE · open
Primary chat channel in TH. Open with PDPA overlay · LINE Official Account terms apply.
🇹🇭 Thailand · V3 chat · open
Patient-initiated · PDPA contract basis · 30d retention default · TH-language fallback at A4.translation review.
🇯🇵 Japan · Email · APPI + Specified Email Act
Required: opt-in + sender identification + reply-to functioning · auto-block if missing.
🇯🇵 Japan · LINE · primary
LINE is the dominant chat channel in JP · Official Account terms · APPI overlay applies for any data export.
🇯🇵 Japan · V3 chat · APPI scoped
Patient-initiated · APPI contract basis · sensitive-data handling rules apply for medical context.
🇸🇦 Saudi Arabia · Email · MoH + advertising restrictions
Saudi MoH restricts unsolicited medical advertising. Enforcement: require A4 review for every send · cannot self-serve. AR-language required + Hijri date format preferred.
🇸🇦 Saudi Arabia · WhatsApp · MoH rules
Allowed for opted-in patients only · MoH advertising rules apply · prefer AR · A4.editorial gates medical claims.
🇸🇦 Saudi Arabia · V3 chat · MoH rules
Patient-initiated · MoH advertising overlay · A4.editorial gates outbound provider responses with claims.
Test sender · simulate gate decisionA2 + A4 + PR2
Paste sample copy + recipient country → instant gate decision. No actual send.
✗ Last test: BLOCKED · 3 violations
"guaranteed" banned phrase (US/TH/SA) · medical-claim "99% success" missing citation · AI-disclosure line missing · would also fail TCPA opt-in check.
AI-disclosure policyH L96
All AI-generated marketing copy auto-tagged with disclosure line per Google E-E-A-T + emerging EU AI Act + USA FTC guidance.
AI-disclosure required globallyenforced
Disclosure language"This message includes AI-assisted content."
Per-locale translation12 locales · Tolgee verified
Auto-append on sendenabled
PR2 can overrideno · A4 senior only

Audit log · rule edits

All compliance-rule edits logged with who/when/why · two-person sign-off required.
Timestamp
Rule edit
Editor
Co-signer
Outcome
2026-04-22 11:14
Quarterly rule review · all 48 cells
A2.alex_p
legal · K. Hanley
approved
›
2026-04-12 09:48
Add: India DPDP 2023 · email + WhatsApp
A4.senior · niran_t
legal · K. Hanley
approved
›
2026-03-28 14:22
Tighten: SA email auto-block → A4 review-required
A4.senior · sarah_k
A2.alex_p + legal
tightened
›
2026-02-14 08:30
Enforce: AI-disclosure globally · all channels
A2.alex_p
A4.senior · sarah_k
enforced
›
Provider-side compliance gate (PR2)
PR2 channel-config view · auto-block warnings + AI-disclosure auto-append at send · cannot bypass without A4 review.
📋 SCENARIO A1 product admin maintains the tier definitions (Starter / Verified / Featured / Trusted+ / PR5) — feature flags per tier, price points, eligibility rules. Tier-edit changes ship behind a paired-review flag and trigger a downstream provider notification job.
63Tier definitions (DB-managed CRUD)
/dashboard/admin/tiers

Tier definitions

A1 Platform Operator · DB-managed CRUD · zero-deploy edits per subscription.md:30
✓
Wedge I live · Tier definitions DB-managed (per subscription.md:30) · A1 sub-permission required · zero-deploy edits
⛨ A1 Platform Operator · sub-permissions
tier-create · tier-edit · tier-disable · tier-grandfather-existing-subs · tier-migrate-providers — each individually grantable. Edits invalidate Redis entitlement cache · audit-logged.

Active tiers · 5 configured

Drag handle to reorder for marketing pages (per subscription.md:36).
Slug · name
Monthly fee
Commission
Ad credit
Ranking ×
Status
Subs
⋮⋮
free
Free
$0
12%
$0
1.0×
active
876
⋮⋮
pro
Pro
$99
10%
$50/mo
1.1×
active
216
⋮⋮
premium
Premium
$299
8%
$200/mo
1.25×
active
84
⋮⋮
enterprise
Enterprise
$1,200+
5%
custom
1.5×
active
24
⋮⋮
wellness_premium
Wellness Premium · piloted Q3
$199
9%
$120/mo
1.15×
pilot
3

Edit drawer · Pro tier (preview)

All 13 tier fields per subscription.md:42-58 · Save button invalidates Redis cache + re-evaluates entitlements for active subscribers.
Tier metadata · 13 fields
Slug
Name
Description
Monthly fee
Currency
Billing period
Commission rate
Ranking boost
Ad credit / mo
Payout cadence
Support SLA (hours)
Display order
Entitlements · 14 boolean flags · subscription.md L60-78
Quotas · 8 numeric · subscription.md L81-87
max_services
max_staff_records
max_packages
max_promotions
max_team_members
max_active_campaigns
max_drip_sequences
max_outbound_msg / mo

Tier transition scheduler

Schedule price / quota changes that apply to new subscriptions only · existing grandfathered (per subscription.md:151).
Pro tier · price change scheduledeffective 2026-09-01
Pro tier: $99 → $129 · applies to new subscriptions only · existing 216 subscribers grandfathered. Transition email auto-sent 14 days before to new-prospect waitlist.

Migrate providers between tiers

Pending bulk migration · Wellness Premium → Premium
3 providers on deprecated wellness_premium tier. Migration applies proration · sends email notification · audit-logged. Preview audit-log entry before commit.
Provider
From → To
Proration
Effective
Banyan Tree Wellness
wellness_premium → premium
+$32 credit
2026-05-15
Chiva-Som
wellness_premium → premium
+$28 credit
2026-05-15
Kamalaya Koh Samui
wellness_premium → premium
+$41 credit
2026-05-15
Disable tier flow
Disabling Pro: 47 active subscribers stay on Pro · new subscriptions blocked · undisable anytime · type-the-name guard (must type "pro" exactly to confirm). Audit-logged · Redis cache invalidation · marketing pages auto-hide tier card.
📋 SCENARIO A1 monetisation admin manages ad inventory — slot definitions, reserve prices, tier eligibility, and pacing caps per surface. Reserve-price changes require an A2 simulation review against historical bids before they go live.
64Ad inventory · slot reserve prices + tier eligibility
/dashboard/admin/ads/inventory

Ad inventory

A1 configures inventory · reserve prices · tier eligibility · bid floors per advertisement.md:31-67 + L177
✓
Wedge I live · Phase 1 inventory: Featured Provider + Sponsored search + Display banner · Phase 2 (auction + patient-meta targeting) Wedge J/K

Inventory types · 4 active

Featured Provider carouselflat-rate
Top-of-city carousel · 6 slots/city · highest-trust providers.
Capacity6 slots / city / day
Pricing modelflat-rate $150/day
Reserve price$120/day
Eligible tiersPro · Premium · Enterprise
Top-of-search Sponsored ResultsCPC auction
Top-3 per query · Discovery primitive · 30% cap enforced.
Capacitytop 3 per query
Pricing modelCPC second-price
Reserve price$0.35 floor
Eligible tiersPro · Premium · Enterprise
Procedure-page sidebar Featuredflat-rate
1 slot per procedure-detail page · contextual relevance.
Capacity1 slot / procedure
Pricing modelflat-rate $80/wk
Reserve price$60/wk
Eligible tiersPremium · Enterprise
Display bannerCPM auction
Editorial pages · CPM auction with classifier review.
Capacity3 slots / page
Pricing modelCPM second-price
Reserve price$4 CPM floor
Eligible tiersPro · Premium · Enterprise

Per-slot configuration · Featured Provider (preview)

Pricing model
Reserve price (USD/day)
Bid floor for new providers
Prevents zero-bid abuse · advertisement.md L179.
Tier eligibility
Disclosure copy (per locale · advertisement.md L50)

City × category eligibility · 6×6 matrix

Click cell to toggle · Open / Reserved-for-tier-X / Disabled.
Dental
Cardio
Ortho
Cosmetic
Fertility
Wellness
Bangkok
Open
Open
Open
Premium+
Open
Open
Chiang Mai
Open
Open
Open
Enterprise
Disabled
Open
Phuket
Open
Open
Open
Disabled
Open
Disabled
Singapore
Open
Open
Open
Open
Open
Open
Seoul
Open
Disabled
Open
Enterprise
Open
Open
Kuala Lumpur
Open
Open
Disabled
Open
Disabled
Open

Reserve price scheduler

Bangkok Dental · reserve price increase scheduledeffective 2026-07-01
Reserve price $150 → $180 · auto-notify active bidders 7 days before · grandfather rule per subscription.md:151 pattern.

Sponsored search · per-keyword × per-location bid floors

Prevents zero-bid abuse from new providers (per advertisement.md:179).
Keyword
Location
Floor
7d activity
Dental implant
Bangkok
$0.35
2,140 imps
›
Rhinoplasty
Chiang Mai
$0.42
1,820 imps
›
Knee replacement
Phuket
$0.28
980 imps
›
Cosmetic surgery
Seoul
$0.55
3,420 imps
›
Suspicious bid patterns flagged by A2 classifier
View detection rules and ad-fraud monitor.
📋 SCENARIO A4 ad-policy moderator reviews submitted ad creatives against the platform policy — medical claims, before/after consent, regulated-keyword filters, and country overlays. Approved creatives enter the auction; rejected ones return to the provider with a fix-it checklist.
65A4 · Ad-policy creative review queue
/dashboard/admin/ads/review

A4 ad-policy review queue

A4.ad-policy sub-tier (NEW · first-time activation) · A2 classifier auto-flags · A4 reviews edge cases · 4h business-hours SLA
✓
Wedge I live · A4 ad-policy sub-tier active · A2 classifier auto-flags · A4 reviews edge cases · backlog SLA: 4h business hours
Queue health · 47 pending · oldest 1h 23m
Target SLA 4h · A4 staffing: 3 active reviewers · capacity OK · risk amber if oldest > 3h · red if > 4h.
amber
A2 CONFIDENCE: >90% · 14 70-90% · 22 <70% · 11 COUNTRY: US EU TH global INVENTORY: Featured Sponsored Display

Pending review · 10 of 47

Provider · campaign
Creative
Inventory
A2 flag · reason
Age
BNH Hospital
Rhinoplasty Q3
[creative thumbnail · 728×90 banner]
Display
0.94 "guaranteed cure" phrase detected
1h 23m
›
Bumrungrad
Dental implant 2026
[creative thumbnail · sponsored search]
Sponsored
0.78 unverified procedure efficacy claim
52m
›
Bangkok Hospital Phuket
Knee replacement
[creative thumbnail · 300×250]
Display
0.82 missing destination disclosure
38m
›
Chiva-Som
Wellness retreat
[creative thumbnail · feat. provider card]
Featured
0.71 AI-generated image · disclosure label missing
28m
›
Banyan Tree
Spa Q3 promo
[creative thumbnail · 970×250]
Display
0.74 before/after image · consent verification needed
22m
›
HMO Asia
Health checkup
[creative thumbnail · sponsored result]
Sponsored
0.76 price-comparison claim unverified
18m
›
Bangkok Hospital
Cardio campaign
[creative thumbnail · 728×90]
Display
0.91 mortality claim · TH MOH approval reqd
14m
›
Anadolu
IVF Q3
[creative thumbnail · feat. carousel]
Featured
0.83 success-rate claim · EU MDR Art 7
11m
›
Asian Hospital
Cosmetic Q3
[creative thumbnail · sponsored result]
Sponsored
0.72 testimonial · provider attribution missing
7m
›
Mount Elizabeth
Oncology campaign
[creative thumbnail · 300×600]
Display
0.96 cancer-cure claim · multi-jurisdictional
3m
›

Review drawer · BNH Hospital (preview)

BNH Hospital · "Rhinoplasty Q3" · 728×90 display banner
A2 classifier flagged · 0.94 confidence · medical-claim-without-evidence + treatment-substitution
CRITICAL · medical claim
Creative preview · A2-flagged spans highlighted
[728×90 display banner thumbnail]

BNH Hospital · world-class rhinoplasty · guaranteed natural results at 100% safety rate · Book consult →

Classifier matched: medical-claim-without-evidence + guarantee-language · per advertisement.md:171 banned-phrases list.
Provider history
2 prior approvals1 prior rejectionPremium tier
A4 decision
Decision rationale (audit-logged)

Country-rule reference

US · FTC
Claims must be substantiated · §255 endorsements rule · disclosure required for sponsored content. Last updated 2026-04.
EU · MDR Art 7
Prohibition on attributing functions/properties not present · therapeutic-claim restriction · multilingual disclosure required. Last updated 2026-03.
TH · MOH
Medical advertising must be MOH-approved · facility-license display required · banned-phrase list (Thai). Last updated 2026-02.
A2 classifier confidence threshold
Current: 0.70 · history of FP / FN trade-off.
Current threshold0.70
7d false-positive rate2.1%
7d false-negative rate0.4%
AI-content disclosure rule
Per EU AI Act + FTC guidance · creative must label AI-generated text/image with "[AI-assisted]".
Rejected creative campaigns auto-paused
Auction surface and classifier model card.
📋 SCENARIO A1 finance admin configures ad-spend billing — auto top-up rules, PR1 (small-clinic) safety thresholds, and dispute-hold workflow. Failed top-ups pause active campaigns and trigger a provider-side notification with one-click resolve.
66Ad-spend billing · auto top-up + PR1 thresholds
/dashboard/admin/ads/billing

Ad-spend billing

A1 sees aggregated billing · failed auto-recharge alerts · PR1-pending top-ups · Stripe / PayPal / Kbank reconciliation per advertisement.md:111-128
✓
Wedge I live · Unified billing · ad spend deducts from credit pool · auto top-up + PR1 approval gates active

30-day billing summary

Total ad spend
$84k
↑ 18% vs prior 30d
Auto-recharge volume
$32k
38% of total · healthy
Failed-recharge incidents
7
2 SLA risk

Failed auto-recharge queue · 7 active

Campaigns auto-paused on recharge failure · provider notified · A1 escalates if > 24h.
Provider
Amount
Reason
Campaigns paused
ETA
Bangkok Hospital Hua Hin
$500
card declined
2 campaigns paused
2h
Asian Hospital
$1,200
FX limit hit
1 campaign paused
4h
Banyan Tree
$300
card expired
1 campaign paused
1h
Anadolu
$800
card declined · 26h
3 campaigns paused
⚠ SLA

PR1-pending top-ups (manual top-ups > $1k)

Threshold per advertisement.md:128 · PR1 sign-off required for high-value manual adds.
Provider
Amount
PR1 reviewer
Pending
SLA
Bumrungrad
$1,500
PR1 Sarah Khan
4h ago
OK
Bangkok Hospital Phuket
$2,000
PR1 Raj M.
18h ago
⚠ approaching
BNH Hospital
$1,200
PR1 Somchai W.
2h ago
OK

Stripe / PayPal / Kbank · daily settlement reconciliation

Per-rail settlement · ad-credit deductions reconciled against provider payment events (per advertisement.md:111-126).
Rail
Daily settlement
7d total
Status
Stripe
$14k
$92k
reconciled
PayPal
$3k
$22k
reconciled
Kbank
฿180k
฿1.2M
2 pending FX

Refund / dispute handler · 4 awaiting A1 review

Provider complaints about ad-spend billing · A2 forensic supports investigation.
Bumrungrad disputes $200 ad spend
Claim: bot traffic · A2 forensic in progress · 28% of impressions flagged · partial refund proposed
in progress
Anadolu disputes $80 ad spend
Claim: campaign paused but billing continued · investigating · likely settlement timing
queued
Per-provider invoice history + ad-fraud detection
PR2 invoice surface and A2 fraud rules.
📋 SCENARIO A1 leadership reviews subscription health — trial-to-paid conversion, gross/net churn, MRR by archetype, LTV by tier. Cohort drops trigger a join with A3 support to identify root cause (onboarding, value, pricing) and a save-play assignment.
67Subscription analytics · conversion · churn · LTV
/dashboard/admin/subscription/analytics

Subscription analytics

A2 Platform Analyst · tier health · Free→Pro→Premium→Enterprise conversion · churn · LTV per subscription.md:161
✓
Wedge I live · Subscription analytics · A2 owns dashboard · Free→Pro conversion is the leading indicator for monetization PMF

Tier distribution · 1,203 active subscribers

Active subscribers per tier
Donut · sorted by tier index · Free is volume base · Premium+Enterprise drive revenue.
1,203
active subs
Free876 · 73%
Pro216 · 18%
Premium84 · 7%
Enterprise24 · 2%
4-stage conversion funnel · 30 days
Free → Pro is the PMF leading indicator.
Free signups1,200/mo
→ Pro upgrades47/mo · 3.9% conv
→ Premium upgrades12/mo · 25% of Pro
→ Enterprise upgrades3/mo · 25% of Premium

Retention per tier · 12-month series

Churn = inverse of retention · monitored monthly · enterprise-grade retention is non-negotiable per wedge spec risk row 1.
Free retention 12mo avg
82%
expected · low-friction
Pro retention
91%
healthy
Premium retention
96%
strong
Enterprise retention
98%
excellent
Free Pro Premium Enterprise

LTV per tier · ARPU × avg lifespan

Tier
Sub fee
Avg commission
ARPU/mo
Avg lifespan
LTV
Free
$0
$42/mo (12% rate)
$42
14 mo
$588
Pro
$99
$84/mo (10% rate)
$183
12 mo
$2,196
Premium
$299
$84/mo (8% rate)
$383
24 mo
$9,192
Enterprise
$1,200+
custom
$1,200+
36 mo
$43,200+

Downgrade-reason analytics · 90 days

Captured at downgrade flow (per pr65) · informs pricing + feature-fit per wedge spec risk row 1 mitigation.
"Features I don't use"
41%
"Too expensive"
28%
"Switching to competitor"
12%
"Other / no reason"
19%
Top reason 41% "features I don't use" suggests Pro tier may bundle too many features for early-stage providers · A2 evaluating thinner Pro variant.

Tier-change events log · last 50

Provider
From → To
Proration
Timestamp
Reason
Bumrungrad
Pro → Premium
+$112 prorate
2026-05-09 14:22
Hit campaign cap
Asian Hospital
Premium → Pro
−$84 credit
2026-05-09 11:08
Features unused
Banyan Tree
Free → Pro
+$99
2026-05-09 09:34
Wanted ad credits
Anadolu
Pro → Enterprise
+$1,101 prorate
2026-05-08 16:50
CSM dedication
HMO Asia
Pro → Free
−$66 credit
2026-05-08 13:15
Too expensive
Auction-collusion + ad-fraud monitor · 7d
Cross-link to a62 for live data.
Auction-classifier alerts (7d)0
Ad-fraud incidents (7d)2 under investigation
Bot-traffic flagged impressions1.4%
Last classifier retrain2026-05-04
Tier-feature-parity drift monitor · 90d
Per wedge spec risk row 6 · ensures entitlement flags match tier definitions.
Audits run (90d)12
Issues found1 resolved
Last incidentpeer-benchmarks accidentally enabled on Free for 3 days
Resolutionrolled back · A1 governance update
Provider-side subscription view + auction surface
PR1 sees own tier · A1 sees aggregate.
📋 SCENARIO Admin reviewer (A1) edits the per-archetype verification rubric — what credentials, certificates, and licences are required for cosmetic vs dental vs IVF vs wellness providers. Rubric changes are versioned; existing approved providers re-run when the rubric tightens.
45A1 per-archetype verification rubric workflow
/dashboard/admin/verification/rubrics

Verification rubrics · 6 archetypes · A1 review

A1 Platform Operator · per-archetype rubric + evidence panel + decision affordances · spec J-archetype-expansion.md:64 + Trust J + Compliance per-archetype overlay
✓ Wedge J live · 6 archetypes
✓
Wedge J live · per-archetype verification rubrics · A1 expanded scope · every decision logged to immutable Compliance audit chain
⛨ A1 Platform Operator · per-archetype rubric owner
Healthcare uses medical license + JCI/ISO baseline · Wellness uses operator license + therapist certs · Diagnostic uses lab accreditation · Insurance uses regulator license + solvency · Facilitator uses partner agreements · Pharmacy uses pharmacy license + controlled-substance authorization. Decision affordances: Verify · Request more info · Reject · Escalate to specialist.
Healthcare Wellness Diagnostic Insurance Facilitator Pharmacy
Healthcare rubric · F+G baselinelive since F
Medical license + JCI/ISO baseline · same rubric since F shipped, deepened at G with practitioner credentials.
Medical license (jurisdiction-issued)required
JCI / ISO 9001 / ISO 15189 cross-checkbaseline
Practitioner credential roster (MD/DDS/RN)G-required
Insurance pre-auth registration (where applicable)conditional
Sanctions / OFAC clearrequired
Healthcare evidence panel · Bumrungrad (review #VER-3401)
Uploaded docs + auto-pull feed status + A1 review queue.
Medical license (Thai MoPH)✓ verified · auto-pull 2026-05-08
JCI accreditation cert✓ valid through 2027-09
Practitioner roster (47 MDs)2 pending re-cred
OFAC + sanctions screen✓ clear
Wellness rubric · spa/operator licensinglive since H
Spa/wellness operator license + therapist certs (per J-archetype-expansion.md:28).
Operator license (spa/wellness regulator)required
Therapist certification rosterrequired
Health-and-safety inspection certconditional
Wellness-claim disclaimer registrationrequired
Wellness evidence panel · Banyan Tree Wellness
Uploaded docs + therapist credential refresh.
Operator license (Thai Spa Association)✓ verified
Therapist roster (18 certified)✓ all current
Wellness-claim disclaimer reg✓ on file
Diagnostic rubric · lab accreditationNEW at J
CLIA-equivalent / ISO 15189 / CAP + proficiency-test history + sample-logistics SOP per spec L319.
Lab accreditation (CLIA-equiv / ISO 15189 / CAP)required
Proficiency-test history (CAP-PT or eq.)required · 12mo
Sample-logistics SOP (chain of custody)required
Pathologist roster + sign-off authorityrequired
Diagnostic evidence panel · BHL Labs (review #VER-3422)
Auto-pull from accreditation feeds + proficiency-test data source.
ISO 15189 accreditation✓ valid through 2026-12
CAP proficiency-test (12mo)⚠ 1 panel scored 89% (threshold 90%)
Sample-logistics SOP v2.4✓ filed
Pathologist roster (4 board-certified)✓ all current
Insurance rubric · regulator licenseNEW at J
Insurance regulator license + solvency disclosure + claim-handling history + sanctions/AML clear per spec L320.
Insurance regulator license (per jurisdiction)required
Solvency disclosure (per regulator threshold)required
Claim-handling history (12mo)required
Sanctions / AML screen (Refinitiv / ComplyAdv)required
Insurance evidence panel · Pacific Cross Health (#VER-3431)
Auto-pull from regulator + sanctions feed.
SEC PH license✓ verified
Solvency margin (latest filing)✓ 218% · regulator threshold 100%
Claim-handling 12mo (median 14d)✓ healthy
Sanctions / AML (Refinitiv pull)✓ clear
Facilitator rubric · partner networkNEW at J
Partner-network signed agreements + complaint history + consumer-protection rule clearance per spec L321.
Partner-provider signed agreementsrequired
Complaint history (12mo · Better Business / equivalent)required
Consumer-protection rule clearancerequired
Margin-transparency disclosure (where applicable)jurisdiction-conditional
Facilitator evidence panel · MedJourney Asia (#VER-3441)
Uploaded agreements + complaint feed cross-check.
Partner agreements (12 providers signed)✓ verified
Complaint history (12mo · 2 minor)✓ within tolerance
Consumer-protection clearance (PH + TH)✓ clear
Margin-transparency (Australia route)⚠ disclosure missing
Pharmacy rubric · drug-authenticityNEW at J
Pharmacy license + controlled-substance authorization + drug-authenticity feed handshake per spec L322.
Pharmacy license (jurisdiction)required
Controlled-substance authorization (per Schedule)required
Drug-authenticity feed handshake (anti-counterfeit)required
Pharmacist roster + license cross-checkrequired
Pharmacy evidence panel · Watson Pharmacy TH (#VER-3451)
Auto-pull from authority + counterfeit feed.
Pharmacy license (Thai FDA)✓ verified
Controlled-substance auth (Schedule II-V)⚠ Schedule II not authorized · escalate
Drug-authenticity handshake (FDA verify)✓ active
Pharmacist roster (8 licensed)✓ verified

Per-archetype reason library · A1 reject / request-more-info reasons

A1 picks reason from per-archetype library when rejecting or requesting more info · standardised wording feeds Compliance audit chain.
Healthcare
License expired · JCI lapsed · practitioner cred missing · OFAC hit
Wellness
Operator license expired · therapist cert lapsed · disclaimer reg missing
Diagnostic
Accreditation expired · CAP-PT below threshold · SOP missing · pathologist roster gap
Insurance
Regulator license expired · solvency below threshold · sanctions hit · claim-handling outlier
Facilitator
Partner agreements missing · complaint history elevated · margin disclosure missing
Pharmacy
License expired · Schedule auth missing · authenticity feed not active · pharmacist roster gap

Audit log · last 8 verification decisions · Compliance chain

Every A1 decision is logged immutably to the Compliance audit chain · per-archetype event-namespacing · feeds ./admin.html#a48 overlays.
Timestamp
Archetype
Provider
Decision
A1 operator
Audit hash
2026-05-09 14:22
Healthcare
Bumrungrad
verify
A1.sarah_k
0x8a4f…91e2
›
2026-05-09 11:08
Diagnostic
BHL Labs
req-info
A1.alex_p
0x2c11…7f80
›
2026-05-09 09:34
Insurance
Pacific Cross Health
verify
A1.sarah_k
0xb8d3…04a1
›
2026-05-08 16:50
Pharmacy
Watson Pharmacy TH
escalate
A1.alex_p
0xff19…b22d
›
2026-05-08 13:15
Facilitator
MedJourney Asia
req-info
A1.sarah_k
0x66ac…7e91
›
2026-05-08 10:01
Wellness
Banyan Tree Wellness
verify
A1.alex_p
0x44f7…2918
›
Onboarding handoff · per-archetype provider flows
Once verified, A1 hand-off lands the provider in the per-archetype onboarding screens.
📋 SCENARIO A2 analyst monitors platform-wide KPIs broken down by archetype — CAC, conversion rate, take rate, retention, ad ROAS. Archetype-level anomalies feed the weekly product council and inform the next Wedge feature prioritisation.
46A2 per-archetype KPI dashboards · platform-wide
/dashboard/admin/insight/archetypes

Platform KPIs · 6 archetypes · cross-archetype attribution

A2 Platform Analyst · per-archetype aggregated dashboards + cross-archetype master + archetype-mix strip · spec J-archetype-expansion.md:65 + Insight J + Observability per-archetype namespacing
✓ Wedge J live · 6 archetypes
✓
Wedge J live · per-archetype KPI dashboards · A2 sees totals across all providers per archetype + cross-archetype attribution feed
⛨ A2 Platform Analyst · per-archetype aggregated insight
Per-archetype KPI dashboards mirror the provider-side (pr92/pr93/pr94/pr95) but A2 sees platform-wide totals. Cross-archetype attribution flows + split-rule audit. Archetype-mix strip informs Subscription tier-pricing per archetype.

Archetype mix · provider count + revenue %

Informs Subscription tier-pricing per-archetype (per J-archetype-expansion.md:143).
Healthcare
847 providers · 58% rev
core archetype
Wellness
312 providers · 14% rev
growing
Diagnostic
94 providers · 9% rev
NEW at J
Insurance
26 providers · 8% rev
NEW at J · attached
Facilitator
19 providers · 6% rev
NEW at J · attribution-heavy
Pharmacy
42 providers · 5% rev
NEW at J · refill recurring
Healthcare Wellness Diagnostic Insurance Facilitator Pharmacy
Total bookings (30d)
12,420
+8% MoM
GMV (30d)
$48.2M
healthy
Avg deposit ratio
22%
deposit transaction shape
Procedure conversion
18%
stable
Total sessions (30d)
9,840
+12% MoM
GMV (30d)
$11.4M
healthy
Pkg attach rate
38%
multi-session uptake
Repeat-visitor 12mo
52%
retention strong
Tests delivered (30d)
6,140
NEW at J
Median turnaround
28h
ranking signal · J
Accuracy (CAP-PT)
98.2%
healthy
Healthcare-attached %
61%
cross-archetype
Policies in force
4,820
NEW at J
Loss ratio
62%
healthy band
Median claim turnaround
14d
ranking signal · J
Denial rate
4.2%
low
Cases closed (30d)
412
NEW at J
Bundle margin avg
11%
healthy
Repeat-customer rate
28%
ranking signal · J
Healthcare-attached %
82%
cross-archetype primary
Orders fulfilled (30d)
8,920
NEW at J
Refill MRR
$182k
true recurring · J L72
Error rate
0.18%
low
Healthcare-attached refill %
44%
cross-archetype

Cross-archetype attribution master · % of revenue cross-attributed

Tracks revenue flowing across archetype boundaries (per J-archetype-expansion.md:33) · split-rule audit feeds Subscription tier reconciliation.
Source archetype
Destination archetype
Attributed GMV (30d)
Split rule
Source share
Dest share
Facilitator
Healthcare
$3.4M
tier-default
12%
88%
Diagnostic
Healthcare
$1.1M
pre-op bundle
100%
0%
Pharmacy
Healthcare
$420k
post-op refill
100%
0%
Insurance
Healthcare
$2.8M
claim disbursement
0%
100%
Facilitator
Wellness
$840k
tier-default
12%
88%

Insight J · unified event-log · per-archetype namespacing

Per J-archetype-expansion.md:92 · every event prefixed with archetype namespace (e.g., diagnostic.test_result_delivered, pharmacy.refill_shipped).
Timestamp
Event
Provider
Archetype
2026-05-09 14:22
diagnostic.test_result_delivered
BHL Labs
Diagnostic
›
2026-05-09 14:18
pharmacy.refill_shipped
Watson Pharmacy TH
Pharmacy
›
2026-05-09 14:12
insurance.claim_paid
Pacific Cross Health
Insurance
›
2026-05-09 14:05
facilitator.bundle_completed
MedJourney Asia
Facilitator
›
2026-05-09 13:58
healthcare.procedure_completed
Bumrungrad
Healthcare
›
2026-05-09 13:45
wellness.session_completed
Banyan Tree Wellness
Wellness
›
Provider-side per-archetype KPIs
A2 sees platform-wide aggregated · provider sees own. Cross-archetype attribution on both sides reconciles via split-rule audit.
📋 SCENARIO A4 policy lead authors per-archetype content rules — what claims are allowed for cosmetic, dental, IVF, oncology, etc. Rule changes deploy to the auto-classifier and refresh the human-review queue's policy hint cards.
47A4 per-archetype content rules
/dashboard/admin/moderation/content-rules

Per-archetype content rules · A4 moderation

A4 Moderator · per-archetype content-rule library + rule-trigger queue + reason library + audit log · spec J-archetype-expansion.md:66 + compliance.md:74
✓ Wedge J live · 6 archetypes
✓
Wedge J live · per-archetype content rules · A4 expanded scope · every moderation logged to Compliance chain
⛨ A4 Moderator · per-archetype content-rule owner
RBAC: edit rule library + review flagged content + approve/reject with reason · cannot disable archetype-mandatory rules · all decisions audit-logged. Rules trigger at PR3 campaign send + provider profile publish + listing edit.
Healthcare Wellness Diagnostic Insurance Facilitator Pharmacy
Healthcare rule library · 14 active
Medical-claim filtering · before/after photo unedited rule · IVF success-rate disclosure.
Medical-claim phrase filter (banned-phrase catalog)mandatory
Before/after photo unedited rulemandatory
IVF / fertility success-rate disclosuremandatory
Outcome-guarantee language banmandatory
Insurance-coverage claim verificationconditional
Healthcare flagged queue · 8 pending
Rule-triggered content with A4 review affordances.
SUB-7841 · Bumrungrad · post-op recovery email⚠ medical-claim phrase
SUB-7845 · Phyathai · IVF success-rate⚠ disclosure missing
SUB-7849 · BNH Hospital · before/after⚠ photo-edit detected
Wellness rule library · 6 active
Wellness-claim disclaimer rules.
Wellness-claim disclaimer ("not medical advice")mandatory
"Cure" / "treats" language banmandatory
Therapist credential surface requiredmandatory
FDA-style disclaimer (US route)conditional
Wellness flagged queue · 2 pending
SUB-7848 · Banyan Tree · "cures stress"⚠ banned phrase
SUB-7850 · Aman Resort · disclaimer missing⚠ disclaimer
Diagnostic rule library · 5 active
Result-interpretation rules · no diagnostic-claim without pathologist sign-off.
No diagnostic-claim without pathologist sign-offmandatory
Test-accuracy claims must cite CAP-PTmandatory
Result-interpretation upsell disclosuremandatory
Genetic-screening informed-consent languageconditional
Diagnostic flagged queue · 3 pending
SUB-7852 · BHL Labs · diagnostic claim⚠ pathologist sign-off missing
SUB-7854 · Quest TH · accuracy claim⚠ CAP-PT cite missing
Insurance rule library · 8 active
Must disclose limitations · cannot use claim-payout history misleadingly · reserved language for "guaranteed" claims.
Coverage-limitation disclosure requiredmandatory
Claim-payout history accuracy + denial-rate contextmandatory
"Guaranteed" reserved language (regulator-specific)mandatory
Pre-existing-condition exclusion claritymandatory
Insurance flagged queue · 4 pending
SUB-7858 · Pacific Cross · "guaranteed payout"⚠ reserved language
SUB-7860 · BUPA · denial-rate misleading⚠ context missing
Facilitator rule library · 5 active
Partner-disclosure mandatory · cannot conceal margin if margin-transparency rule applies.
Partner-provider disclosure mandatorymandatory
Margin-transparency (where jurisdiction applies)conditional · per a48
Bundled-trip itinerary accuracymandatory
Complaint-history surfacing (where required)conditional
Facilitator flagged queue · 2 pending
SUB-7862 · MedJourney · Australia margin⚠ disclosure missing
SUB-7864 · GlobalMed · partner not listed⚠ disclosure
Pharmacy rule library · 10 active
Cannot claim "cures" · controlled-substance marketing prohibited per Schedule · counterfeit-warning required on bulk pricing.
"Cure" claim banmandatory
Controlled-substance marketing ban (Schedule II-IV)mandatory
Counterfeit-warning on bulk pricingmandatory
Rx-required clarity (Rx vs OTC chip)mandatory
Refill-cadence advertising (chronic only)conditional
Pharmacy flagged queue · 5 pending
SUB-7866 · Watson · Schedule II ad⚠ controlled-substance
SUB-7868 · Boots · "cures" copy⚠ banned claim
SUB-7870 · Mercury · counterfeit warning⚠ missing

Per-archetype reason library · A4 reject / approve-with-edits

A4 picks reason from per-archetype library when actioning a flagged item · feeds Compliance audit chain.
Healthcare
Medical-claim phrase · photo-edit detected · disclosure missing · outcome-guarantee
Wellness
"Cures" claim · disclaimer missing · therapist credential not surfaced
Diagnostic
Pathologist sign-off missing · CAP-PT cite missing · interpretation upsell undisclosed
Insurance
Reserved language · denial-rate context missing · limitation undisclosed · pre-existing exclusion unclear
Facilitator
Partner not listed · margin disclosure missing · itinerary inaccurate
Pharmacy
Schedule violation · "cures" claim · counterfeit warning missing · Rx/OTC unclear

Audit log · last 6 moderation decisions · Compliance chain

Timestamp
Archetype
Submission
Decision
A4 moderator
Audit hash
2026-05-09 14:30
Pharmacy
SUB-7866 · Watson Pharmacy
reject
A4.niran
0x77ff…0a31
›
2026-05-09 13:14
Insurance
SUB-7858 · Pacific Cross
reject
A4.sarah_k
0x18bb…f042
›
2026-05-09 11:18
Healthcare
SUB-7841 · Bumrungrad
approve-edits
A4.sarah_k
0x9e02…1da5
›
2026-05-09 10:02
Wellness
SUB-7848 · Banyan Tree
reject
A4.aksu_y
0x6660…cfe7
›
2026-05-09 09:14
Diagnostic
SUB-7852 · BHL Labs
req-info
A4.niran
0x5500…8aa9
›
2026-05-08 16:42
Facilitator
SUB-7862 · MedJourney
req-info
A4.aksu_y
0xae33…7012
›
Wired to A4 review queue + per-archetype provider profiles
Existing A4 review queue absorbs flagged content; reason library standardises wording across archetypes.
📋 SCENARIO A1 regulatory officer maintains the per-archetype regulatory overlay — country-specific rules layered on top of the base policy (e.g. Singapore HSA, Thai FDA, EU MDR). Overlay updates re-validate active campaigns, listings, and onboarding flows in the affected country.
48Per-archetype regulatory overlay config
/dashboard/admin/compliance/overlays

Regulatory overlays · 6 archetypes × jurisdictions matrix

A1 + A2 (Compliance Officer sub-role) · per-archetype × per-jurisdiction overlays · Compliance major extension per J-archetype-expansion.md:36 + Identity external-partner API key mgmt per identity-multitenant.md:49
✓ Wedge J live · 6 archetypes
✓
Wedge J live · per-archetype regulatory overlays · A1 + A2 Compliance-Officer sub-role · sanctions/AML + counterfeit + lab-feed + consumer-protection partners configured per archetype
⛨ Compliance Officer sub-role · per-archetype overlay owner
Each archetype × jurisdiction cell loads its own overlay: Insurance has solvency/AML; Pharmacy has Schedule + counterfeit; Diagnostic has CLIA/ISO/CAP feeds; Facilitator has consumer-protection. RBAC: A1 + A2 + legal counsel · two-person sign-off · audit-logged.

Archetype × jurisdiction matrix · click any cell to load overlay

Green = live · Amber = partial config · Grey = not applicable. 5 archetypes (excluding Healthcare baseline) × 6 jurisdictions.
Archetype
🇺🇸 US
🇪🇺 EU
🇹🇭 TH
🇸🇬 SG
🇸🇦 SA
🇦🇺 AU
Insurance
✓ NAIC
✓ EIOPA
✓ OIC
✓ MAS
⚠ partial
✓ APRA
Pharmacy
✓ DEA
✓ EMA
✓ Thai FDA
✓ HSA
⚠ SFDA
⚠ TGA
Diagnostic
✓ CLIA
✓ ISO 15189
✓ ISO 15189
✓ SAC
⚠ partial
✓ NATA
Facilitator
✓ FTC
✓ UCPD
✓ OCPB
✓ CCCS
— low usage
⚠ ACL margin
Wellness
✓ FDA-style
✓ general
✓ Spa Assoc
✓ general
— low usage
✓ general
🇺🇸 Insurance · NAIC overlay · solvency + sanctions/AML
Solvency margin threshold: 100% RBC · Claim-handling rules: 30d response window · Sanctions/AML provider: Refinitiv (open per J-archetype-expansion.md:132 — pending switch evaluation Q3 vs ComplyAdvantage vs native).
Solvency threshold
100% RBC
Claim-handling window
30d
Sanctions provider
Refinitiv
🇹🇭 Pharmacy · Thai FDA overlay · controlled-substance + counterfeit
Controlled-substance schedules: II-V mapped to Thai PSO. Granularity TBD per J-archetype-expansion.md:125 · current = country-level · per-state evaluation pending. Counterfeit-prevention partner: Thai FDA verify endpoint. International shipping rules deferred per J-archetype-expansion.md:133.
Schedule auth
II-V
Counterfeit partner
Thai FDA verify
Intl. shipping
deferred
🇺🇸 Diagnostic · CLIA overlay · accreditation feeds
CLIA-equivalent feed: CMS QCOR. ISO 15189 cross-check feed: A2LA. Proficiency-test data source: CAP-PT.
CLIA-equiv feed
CMS QCOR
ISO 15189 cross-check
A2LA
Proficiency-test source
CAP-PT
🇦🇺 Facilitator · ACL overlay · margin-transparency mandatory
Australian Consumer Law applies margin-transparency; Facilitator providers must disclose margin to patient. Consumer-protection rules surfaced via ACCC reference + complaint history feed.
Margin-transparency
required
Consumer-protection ref
ACCC
Complaint feed
Better Bus.

Audit log · last 6 overlay changes · Compliance chain

Every overlay change logged · two-person sign-off captured · feeds ./provider.html#pr80/pr82/pr84/pr86 Compliance audit chain.
Timestamp
Archetype × jurisdiction
Change
A1/A2 author
Sign-off
2026-05-09 13:01
Insurance · 🇺🇸 US
Sanctions provider · Refinitiv → ComplyAdv. simulated
A2.alex_p
legal · K.Hanley
›
2026-05-08 11:32
Pharmacy · 🇹🇭 TH
Schedule II auth restricted · 4 providers re-verify required
A1.sarah_k
legal · K.Hanley
›
2026-05-07 09:14
Facilitator · 🇦🇺 AU
Margin-transparency rule turned mandatory
A2.alex_p
legal · K.Hanley
›
2026-05-04 14:52
Diagnostic · 🇸🇬 SG
SAC accreditation feed activated
A1.sarah_k
legal · K.Hanley
›
2026-04-29 10:08
Insurance · 🇪🇺 EU
Solvency threshold raised 100% → 110% (regulator alignment)
A2.alex_p
legal · K.Hanley
›
2026-04-22 16:40
Pharmacy · 🇦🇺 AU
TGA counterfeit-prevention partner pending integration
A1.sarah_k
pending
›
⚠ Open questions · pointers to J-archetype-expansion.md
• Sanctions/AML provider final selection — Refinitiv vs ComplyAdvantage vs native (per J-archetype-expansion.md:132) · 30d simulation in flight
• Pharmacy controlled-substance Schedule granularity — country-level vs per-state (per J-archetype-expansion.md:125) · per-state evaluation queued
• Pharmacy international-shipping rules — deferred (per J-archetype-expansion.md:133) · revisit in K wedge
• Insurance Actuary persona promotion (NOTE.md decision 2 resolved) — confirm tier-pricing per archetype (per J-archetype-expansion.md:143)
Provider-side per-archetype Compliance audit chain
Overlay changes propagate to per-archetype provider Compliance views.
📋 SCENARIO A1 financial-product underwriter reviews a working-capital loan application from a verified provider — bookings cashflow, GMV trend, repayment ratio, and KYC bundle. They approve, decline, or counter; approved loans disburse to the provider's settlement account and accrue against future bookings.
88A1 lending workspace · working-capital underwriting (K.3 Capital)✓ Wedge K · K-spec capability #14
/dashboard/lending

A1 lending · working-capital advances

Provider working-capital underwriting · booking-history threshold · APR config · repayment tracking · Toast Capital benchmark

A1 financial-product compliance K-spec line 105
Active advances
14
Disbursed YTD
$480k
In repayment
12
Default rate
0%
Amber-flagged
2
Bangkok Hospital Phuket · $40k requested
Threshold ✓OFAC ✓Risk 0.42
Booking history 14 mo · avg monthly $11.2k · no payment disputes · sanctions screening clean (a92) · risk score from a91
Silver tier · APR 11%Amount band $25k-$60k36-month repayment
Bumrungrad Bangkok · $75k requested
Threshold ✓OFAC ✓Risk 0.18
Booking history 22 mo · avg monthly $34.8k · 0 disputes · gold-tier provider · > $60k threshold → escalate to A1 lead nick_admin
Gold tier · APR 8%Amount band $60k+Escalation required
HealthCity Manila · $18k requested
Threshold ✗OFAC ✓Risk 0.61
Booking history 4 mo (need 6) · avg monthly $6.4k (need $8k) · auto-rejection candidate · A1 may override with documented justification
Lending decision-support cross-links
Audit chain · WHO / WHAT / WHEN per advance · escalations · sanctions screening · risk score
📋 SCENARIO A2 federated-learning engineer orchestrates a training round across N≥10 consenting providers, with secure aggregation and differential-privacy budget tracking. Round outputs are evaluated against held-out evals before the global model is promoted to production.
89A2 federated learning · model training across N≥10 providers✓ Wedge K · K-spec capability #17
/dashboard/data/federated

Federated learning · model training

Trains across N≥10 providers · per-provider contribution caps · model-update validation · demonstrable improvement vs centralized baseline

A2 federated engineer K-spec line 106
Active models
4
Participating providers
14
Last training round
2h ago
Avg improvement
+4.7%
M-match-2026-q2 · Matching modelDeployed · v3.2
Federated training across 14 providers · last round 2h ago · contribution cap 15% per provider · K-spec acceptance criteria #6 (≥10 participants) ✓
Participants
14 / ≥10 ✓
Contribution cap
15% max / provider
vs centralized baseline
+5.2% accuracy
Demonstrable improvement chart
Federated: ▁▂▃▄▅▆▇█ · accuracy 0.847
Centralized: ▁▁▂▃▃▄▄▅ · accuracy 0.795
Δ +5.2% over 12 rounds
K-spec acceptance: must demonstrate improvement vs centralized baseline before shipping.
⚠ Model-update validation · provider PR-7821 update rejected
Update validated against held-out test set · anomalous loss curve · likely poisoning attempt · audit-logged · provider notified · K-spec line 179 · contribution cap protected aggregate

Provider participation matrix

Provider
Opt-in
Contribution
Updates last 7d
Revenue share
Bumrungrad Bangkok
opted in
14% / 15% cap
7 of 7
$1,840
›
Bangkok Hospital Phuket
opted in
12% / 15% cap
7 of 7
$1,520
›
Asian Heart Institute
opted in
9% / 15% cap
6 of 7
$1,210
›
HealthCity Manila (PR-7821)
suspended
— rejected
0 of 7
$0
›
Architecture Per-provider environment hosts local training · model deltas aggregated centrally · raw data never leaves provider
Privacy-preserving guarantees: differential privacy ε=2 · secure aggregation (multi-party) · per-record k-anonymity ≥5 · K-spec line 147
Federated learning operations
📋 SCENARIO A2 AI governance engineer runs the eval harness on candidate model updates — accuracy, hallucination rate, jailbreak resistance, locale-fairness, and cost per query. Failing evals block deployment; cost regressions trigger the guardrail throttle.
90A2 AI eval harness · governance + cost guardrails✓ Wedge K · K-spec capability #18
/dashboard/data/eval

AI eval harness · governance

Clinical accuracy + bias audit per matching/ranking model · cost guardrails block overspend · A2 + external clinical advisor sign-off · K-spec line 154 + 172

A2 + external clinical advisor
Models in production
6
Last nightly run
04:00 UTC
Regression detected
1
Cost YTD
$14.2k
Bias flags
2
⚠ M-rank-2026-q2 · 3.2% accuracy drop overnightAuto-rollback queued
Likely cause: training data drift · A2 + external clinical advisor paged · A1 sign-off chain triggered (K-spec line 172) · auto-rollback to v2.3 queued pending advisor review
M-match-2026-q2 · MatchingHealthy
Clinical accuracy 0.847 · bias disparity max 4.2% · cost / decision $0.0012 · drift 0.08 · last regression none in 30d
M-rank-2026-q2 · RankingRegression
Clinical accuracy 0.731 (was 0.763 · −3.2%) · bias disparity 6.8% · cost / decision $0.0019 · drift 0.31 · auto-rollback queued
M-outcome-2026-q2 · OutcomeBias flag
Clinical accuracy 0.782 · bias disparity 12.4% (age × procedure intersection) · cost / decision $0.0034 · A2 review queued
M-fraud-2026-q1 · Fraud detectionHealthy
Clinical accuracy 0.916 · bias disparity 2.1% · cost / decision $0.0008 · drift 0.04 · last regression 87d ago
A2 + A1 + external advisor sign-off chain
📋 SCENARIO A2 engineer monitors the composite procedure × provider × patient risk score, opens the calibration view, and updates federated-learning model weights when bias-audit drift exceeds threshold. High-risk bookings auto-route to A1 for clinical-safety pre-clearance.
91A2 risk dashboard · composite procedure × provider × patient✓ Wedge K · K-spec capability #19
/dashboard/data/risk

Composite risk dashboard

Procedure × provider × patient risk score · feeds insurance underwriting (mb06) · A1 lending (a88) · patient transparency · last retrain 7d ago

A2 maintains model 6 inputs
Input 1
Procedure category
e.g. cardiac · IVF · ortho
Input 2
Provider verification tier
bronze / silver / gold (a94)
Input 3
Provider booking history
volume · disputes · outcomes
Input 4
Patient KYC level
basic / verified / passport
Input 5
Patient health profile
comorbidities · age
Input 6
Cross-border jurisdiction
origin × destination
Output channels
→ Insurance underwriting (mb06) → Patient transparency chip (mb06) → A1 lending decision support (a88) → Patient guidance ("flagged risky combination")

Procedure × provider risk · color-coded

Bronze Silver Gold
Cardiac CABG 0.71 high 0.42 med 0.18 low
Cardiac TAVR 0.78 high 0.51 med 0.29 med
IVF cycle 0.31 low 0.19 low 0.12 low
Ortho TKR 0.48 med 0.27 low 0.16 low
Bariatric 0.53 med 0.34 med 0.21 low
Retrain risk modelLast retrain 7d ago · cadence weekly
📋 SCENARIO A1 financial-product compliance officer reviews provider screening results against OFAC + UN sanctions lists. Flags trigger a 48h hold on payment settlement until cleared or rejected; confirmed matches escalate to legal and freeze the provider account.
92A1 sanctions / AML screening · OFAC + UN list reconciliation✓ Wedge K · K-spec line 125
/dashboard/compliance/sanctions

Sanctions / AML screening

Provider screening · OFAC + UN list reconciliation · settlement-hold workflow · K.2 Insurance + K.3 Capital prerequisite · trust-and-safety.md line 49

A1 financial-product compliance Provider: Refinitiv · evaluating
Providers screened
234
Flagged providers
2
Patients flagged
0
Re-screening cadence
30d auto

Provider screening table

Provider
OFAC
UN list
Last screened
Flag reason
Resolution
Bumrungrad Bangkok
Clean
Clean
2026-04-18
—
OK
›
Bangkok Hospital Phuket
Clean
Clean
2026-04-15
—
OK
›
MedTrust Lebanon
Partial
Clean
2026-05-08
Owner name partial match · false-positive likely
Review
›
CrimsonHealth Dubai
Hit
Hit
2026-05-09
Beneficial owner sanctioned
Hold
›
⚠ CrimsonHealth Dubai · settlement-hold detailA1 resolution required
Hit detail: beneficial owner Mohammed K. matches UN sanctions list (added 2024-08-12 · sanctions order #UN-2024-1428) · OFAC SDN list secondary match
Settlement hold
$48k pending
SLA window
48h · A1 review
A1 resolution log
3 entries
Re-screening cadence configuration
Auto re-screen every30 days
Trigger re-screen on payment >$10,000
Trigger re-screen on advance applicationenabled
Settlement hold workflow SLA48h
Patient AML
Patients screened only at high-value bookings (cumulative > $25k) per K-spec · 12 patients screened YTD · 0 flagged · K.4 governance
Provider integration · Refinitiv vs ComplyAdvantage vs native — current Refinitiv · A1 evaluating · K-spec line 168
📋 SCENARIO A1 research-product admin manages provider and patient enrollment in the de-identified clinical research data product, oversees IRB approvals, and configures the revenue-share split. Enrollment closes automatically once consent thresholds or jurisdiction caps are hit.
93A1 clinical research data product · enrollment + revenue-share admin✓ Wedge K · K-spec capability #16
/dashboard/data/research

Clinical research network

Partner contracts · provider + patient enrollment · anonymization pipeline · revenue distribution · cryptographic deletion proof · K-spec line 178

A1 governs A2 maintains pipeline
Partner contracts
8
Enrolled providers
47
Enrolled patients
2,340
YTD revenue
$342k
Revocations YTD
14

Partner contracts · revenue split provider 60% / patient 5% / ArokaGO 35% (per mb17)

Partner
Cohort
Contract value
Revenue split
End date
Pfizer
Oncology cohort (post-market)
$120k
60 / 5 / 35
2027-03-15
›
Medtronic
Orthopedic device outcomes
$95k
60 / 5 / 35
2027-08-22
›
Mayo Clinic
Post-market surveillance
$78k
60 / 5 / 35
2026-11-30
›
5 additional partners
Mixed cohorts
$49k avg
60 / 5 / 35
Various
›
Opt-in count
47
Avg per-provider revenue
$4.4k
Withdrawn this year
3
Patient revenue surface
📋 SCENARIO Admin reviewer (A1) audits tier-promotion submissions — providers climbing from Verified to Trusted+ submit additional credentials, on-site evidence, and quality outcomes. Approvals unlock new placement weights; denials return with a remediation rubric and a re-apply window.
94Tiered verification climb · A1 audits tier-promotion submissions✓ Wedge K · K-spec capabilities #20-21
/dashboard/trust/tier-climb

Tier promotion queue

Providers submit new evidence to climb bronze → silver → gold → flagship · A1 audits · scoped per-procedure verification · K.5 Stretch Trust

A1 audits K-spec #20-21
Pending climb requests
6
Demotion reviews
2
Flagship providers
3
Avg promotion SLA
5.2 days
Bangkok Hospital Phuket · bronze → silverPending
Evidence: JCI accreditation cert (2026) · 78 verified bookings (need 50+) · 6mo+ activity ✓ · all silver criteria met
JCI ✓ 78 / 50 bookings activity ✓
Bumrungrad Bangkok · silver → goldPending
Evidence: cardiac specialty board cert · oncology specialty board cert · 234 verified bookings (need 200+) · outcome-tracked ✓
Cardiac board ✓ Onco board ✓ Ortho board pending 234 / 200 bookings

Tier criteria · admin-tunable · K-spec line 62

Tier
Verification
Bookings
Specialty certs
Activity
Bronze
Binary verification
—
—
6mo+
Silver
+ JCI accreditation
50+ verified
—
6mo+
Gold
+ JCI
200+ verified
per-procedure board cert
outcome-tracked
Flagship
+ JCI + ISO + audit
500+ verified
multi-specialty board
3y outcome-tracked
Scoped verification · K-spec capability #21
Per-procedure verification: provider verified for IVF / not verified for cardiac · admin-tunable per-procedure · prevents tier inflation across unverified specialties
Bumrungrad Bangkok
cardiac ✓onco ✓ortho ?
IVF Kuala Lumpur
IVF ✓cardiac ✗
Bangkok Hospital Phuket
cardiac ✓onco ✓ortho ✓
Auto-demotion criteria
Tier auto-demotion if booking dispute rate > 5% · A1 reviews before demotion · provider notified · audit-logged
Patient-facing impact
Promotion updates tier badge across patient passport · search filters · doctor portal
📋 SCENARIO A4 moderator reviews individual doctor personal-brand portals — bio, credentials, photos, video introductions — for accuracy and policy compliance separate from the parent provider profile. Conflicts (e.g. doctor claims credential the provider didn't verify) escalate to A1.
95Doctor personal-brand portal admin · A4 review✓ Wedge K · K-spec capability #25
/dashboard/trust/doctor-portals

Doctor brand portals · moderation

A4 (with PR4) reviews verified outcomes + before/afters with consent + video case studies · syndication controls · doctor-specific rating aggregation

A4 review + PR4 sign-off
Active portals
47
Pending review
12
Flagged
3
Doctor-specific ratings
3,840
Dr. Pongsak C. · Bumrungrad Bangkok
cardiac surgery · 234 doctor-specific ratings · avg 4.8 · feeds portal trust score (from mb10 / p35)
Active PR4 verified
Verified outcomes
142
Before/after pairs
38 · all consented
Video case studies
12
Syndication
IG · TT · YT
Dr. Anil K. · Bangkok Hospital Phuket
orthopedic surgery · 187 doctor-specific ratings · avg 4.6 · 1 video case study pending review
Pending review PR4 verified
Video case study moderation queue · 1 video
"Robotic-assisted knee replacement · post-op month 3" · 4:32 duration · transcript review pending · medical-claim flagging: 1 phrase ("guaranteed mobility") flagged for A4 + PR4 review before publication
Dr. Mei L. · Asian Heart Institute
cardiology · 92 doctor-specific ratings · avg 4.2 · 2 before/after pairs flagged · consent revocation reported
Flagged
Outcome-photo verification flag
Patient revoked consent for 2 before/after pairs (post-publication) · cryptographic deletion proof requested · auto-takedown enforced within 24h · A4 sample audit confirms compliance
Syndication controls · K Growth integration
Instagram
38 portals active
142k impressions · 3.4k saves · 280 profile clicks (last 30d)
TikTok
22 portals active
418k views · 8.1k saves · 412 profile clicks (last 30d)
YouTube
14 portals active
89k views · 2.1k saves · 174 profile clicks (last 30d)
Doctor brand portal cross-surfaces