boolanga
×
hungerstation
RFP-AMB-001 · Volume A — Technical · May 4, 2026

Ambassadors Program
Technical Proposal

All five TSD modules. AI visual compliance scoring — live in MENA. KSA-resident hosting. September 2026 go-live.

5
TSD modules
60
Ambassadors
17 wks
To go-live
AI
Bundled · live MENA
Platform architecture

PWA delivery — and why native iOS/Android cannot go live by July.

The TSD specifies Native iOS and Android. Boolanga delivers a Progressive Web Application. This section addresses the gap directly.

✗ Native iOS + Android — 6+ months minimum
2,000+ hours. Separate Swift and Kotlin codebases. 6-month build at any realistic team size — not 8 weeks.
App Store approval. Apple review 1–7 days per submission, rejectable. One rejection delays go-live by weeks outside any vendor's control.
Double QA surface. Every bug fix must be built and tested separately on both platforms.
Device dependency. Ambassador devices must meet minimum OS version requirements before day one.
✓ Boolanga PWA — every operational requirement met
✓ GPS double-capture + KSA bounding box enforcement
✓ AES-256 offline storage — 200 checks, auto-sync
✓ Live photo enforcement — EXIF 60s, Laplacian, MD5
✓ Instant updates to all 60 ambassadors simultaneously
~ Push: full on Android · in-app alerts on iOS
September 2026 full go-live. Native apps available Year 2.
Any vendor promising July for this scope has not costed the build.
Full TSD — 5 modules, RiderCare integration, KSA hosting, pen test, 60-ambassador training — is 17 weeks minimum on the PWA architecture. September 2026 is the fastest credible date.
Interactive prototype — same platform deployed at Talabat

The compliance check — as the ambassador sees it. Try it.

This is the exact same platform delivered live across Talabat's MENA markets, adapted for HungerStation's 8-step flow. Steps 04, 05, and 06 include AI visual scoring. Click through the phone.

→ INTERACTIVE — CLICK THROUGH THE PHONE
Working prototype of the ambassador compliance check. Follow the steps below in order.
01
Search Rider by ID
Ambassador enters the rider ID on the numeric keypad. System loads the rider from WMS — name, 3PL, vehicle, last check date.
Enter any ID, then tap Search →
02
Review Rider Profile + Gear
System shows the rider's gear inventory, expired items, violation history, and last check date. Same data as the MENA markets use today.
Tap "Start check" to begin the inspection
AI
Steps 04/05/06 — AI visual scoring
Photo first. AI scores Steps 04, 05, and 06 from the proof photo automatically — uniform, branding, safety gear. Ambassador reviews and overrides anything wrong.
Tap "Take Photo" — watch AI spinner and results appear. Click Disagree to override any item.
08
Submit — one tap at the end
Single submit. GPS re-captured at submit. Both GPS points stored. Discrepancy >200m auto-flags to Crystal. Live override count in footer.
Tap "Submit Check" to complete
AI compliance module

Visual scoring from one photo. Bundled — no extra charge.

Steps 04, 05, and 06 are visual evaluations. AI replaces the manual checklist with instant analysis from the proof photo already required in Step 07. Non-visual items stay manual.

WITHOUT AI
10–12
minutes per check
WITH AI
~6
minutes per check
★ Bundled — no separate AI charge
AI infrastructure, model maintenance, and retraining as HungerStation updates gear standards are inside the monthly fee. No AI line item, no usage cost.
1
Ambassador takes proof photo
Same photo required by the TSD in Step 07. No extra action from the ambassador.
2
App calls Helmi AI inference API
Photo + metadata sent. No new hardware required on the ambassador device.
3
AI scores in under 3 seconds
Uniform items, branding, helmet, safety gear — each scored individually with confidence level.
4
Pass/fail + violation codes returned
Ambassador reviews and overrides any item where the AI was wrong.
5
Fallback if API unavailable
Manual checklist activates automatically. No check is lost and no ambassador is stuck.
Complete TSD coverage

All 5 modules. Every requirement. Itemised.

Everything below is in scope. Nothing is a roadmap item or separate charge.

AAmbassador Mobile Application12 items
8-step compliance check flow — sequential, step-level validation, no skip permitted
Rider search by ID linked to HungerStation WMS — name, 3PL, vehicle type, last check. Not found: FTR flag.
Live photo enforcement: camera-only, EXIF 60-second window, Laplacian blur, MD5 dedup, anti-screenshot
GPS captured at photo shutter AND at submit — both stored. >200m discrepancy auto-flags for Crystal.
KSA GPS bounding box (lat 16.0–32.5, lon 36.0–56.5) — out-of-range flags check and notifies Team Leader
Impossible movement detection — two checks >5km apart within 3 minutes auto-flagged
AES-256 offline storage — 200 checks capacity, auto-sync on reconnect, timestamps immutable
Daily check counter X/60, progress bar (red→amber→green), push notification at 50 checks
Arabic RTL UI — full mobile app. Boolanga provides all Arabic translations.
Ambassador mobile dashboard: counter, KPI score and tier, pending coaching tasks, last 10 checks
AI visual scoring Steps 04/05/06 — uniform, branding, safety gear from single proof photoAI
Phase 2 city config — Mecca, Medina, Eastern Province, Abha, Al Ahsa, QP — config only, no re-engineering
BIncident & Emergency Response10 items
RiderCare bidirectional integration (8001111717) — webhook + API, real-time event-driven
Live dispatch map — all ambassadors as GPS pins, 30-second refresh, names and availability visible
One-tap dispatch — push to ambassador instantly. 60-second accept/decline window, auto-reoffer.
15-minute on-scene SLA timer visible on agent screen from dispatch time
L1 (serious injury/fatality): 911 required, RiderCare mandatory, HSS Asst Manager + Ops Director same day
L2 (minor accident/breakdown): on-scene ≤15 min, Najm 920002020 if needed, report within 24 hours
L3 (non-cooperation/violation): photo/video documentation, coaching call, Team Leader notified
L1/L2 incident report: rider condition, 2+ photos, GPS auto-captured, Najm reference, 24h deadline
Push escalation to Team Leader and HSS within 5 seconds for L1/L2 incidents
FMS API integration — real-time GPS feed for dispatch map, 30-second refresh
CField Intelligence & Reporting6 items
Voice of the Rider: morale 1–5, satisfaction 1–5, top challenge dropdown + free text — min 1/week enforced
Vendor-Side Issues: vendor, issue type, severity Low/Medium/High, mandatory photo — auto-push on High
Street & Field Reporting: TL-activated, road closure Y/N, congestion 1–5, reroute + map pin
Daily Operational Log: mandatory end-of-shift, auto-filled check count, incident/coaching/FTR fields
All 4 forms count toward ambassador monthly Report Quality KPI score
Non-submission within 30 min of shift end: flag to Team Leader + reduces valid submission count
DKPI Scoring & Incentive Engine13 items
Automated 4-pillar monthly KPI calculator — zero manual intervention
Checks Completed (50 pts): (actual ÷ 1,560) × 50 — proportional, capped at 50
Report Quality (20 pts): ≥95% = 20 pts; each 1% below 95% = −2 pts; below 85% = 0
Proactive Escalations (15 pts): 2+ HSS-validated = 15; 1 = 8; 0 = 0
Violation Detection (15 pts): 5+ = 15; 3–4 = 10; 1–2 = 5; 0 = 0
Deductions: SAR 4/missed check (auto); SAR 15/Crystal-confirmed invalid — never on flag alone
Vehicle incident: 30% base + 50% incentive/month, spread max 3 months — HR + Fleet confirmation required
Minimum payout SAR 0 — deductions cannot result in negative payout
Incentive brackets: Platinum SAR 1,200 / Gold SAR 900 / Silver SAR 650 / Bronze SAR 400 — auto-assigned
TL composite: 60% team avg + 25% eligibility + 15% weekly reports on time — TL read-only
HSS Management Dashboard: cross-city aggregate, city comparison, FTR map, payout projection, alerts
Data export to Excel (.xlsx) + Google Sheets API v4 — Admin and HSS Asst Manager only
HSS Asst Manager manual override with immutable audit log entry per override
ECompliance Audit — Crystal Team10 items
Crystal validation queue: sorted oldest first, per row shows ambassador, city, rider ID, timestamp, GPS, photo, flags
Actions: Approve / Reject (reason mandatory) / Flag — batch approve/reject available
SAR 15 only after Crystal manual confirmation — never automatically on flag alone
48-hour SLA: all submitted checks reviewed within 48 hours
Auto-detection: duplicate GPS within 10m/5min by same ambassador
Auto-detection: MD5 photo hash match across all submissions
Auto-detection: speed anomaly, sequential checks <90 seconds, EXIF age >60s
Coaching call auto-task per violation in Steps 02/04/05/06 — ambassador notified with violation type
Coaching outcome logging: Completed / No answer / Refused / Accepted. 24h non-completion → auto-escalate to TL.
WMS API read-only — on-demand at Step 01, 4-hour cache refresh, offline manual fallback
+Platform Infrastructure & Security12 items
5-role RBAC enforced at API and database level — cross-role access impossible at system level (P0 critical)
SSO with HungerStation IDP + 2FA mandatory for HSS Manager and Admin
AES-256 at rest, TLS 1.3 in transit, signed photo URLs 1-hour expiry, no predictable paths
Immutable audit logs — 2-year retention, all submissions/edits/exports/deductions/overrides
External pen test — all Critical/High resolved before go-live, report to HungerStation CISO
12 mandatory P0/P1 automated test cases validated before go-live authorisation
Push: FCM full on Android · in-app alerts on iOS. L1/L2 delivery confirmation required.
KSA-resident hosting — 100% KSA, PDPL + NDMO compliant, Data Residency Attestation Letter pre-signature
Performance: submission <5s on 4G; dashboard <3s; 3,600+ submissions/day without degradation
Helmi AI API integration — Steps 04/05/06, auto-fallback to manual on API unavailabilityAI
Training: 60 ambassadors, 2 Team Leaders, HSS management — proficiency test before go-live
Arabic RTL implementation — Boolanga provides all translations
Not included in this proposal
Native iOS and Android apps — available as a Year 2 engagement
Direct HR/payroll API — incentive/deduction data exportable to Excel/Sheets for HS payroll
PDPL legal advisory — Boolanga provides KSA hosting certificate; legal consultation not included
Delivery plan

17 weeks. Four stages. September 2026.

Client dependencies required at kickoff: RiderCare API credentials, WMS/FMS API access, HungerStation IDP details, and 10 pilot ambassadors for Stage 4 UAT.

WEEKS 1–4 · STAGE 1
Core App + Infrastructure
UI redesign · Photo enforcement · RBAC · SSO · AES-256 · Push · Arabic RTL · Module A (full check flow, offline, counter, dashboard) · KSA hosting · WMS integration
✓ Core check flow operational
WEEKS 5–8 · STAGE 2
Incident & Emergency
Module B: RiderCare integration · Live dispatch map · L1/L2/L3 workflows · Incident reports · SLA timer · Push escalation · FMS GPS · Pen test commissioned
✓ RiderCare dispatch live
WEEKS 9–12 · STAGE 3
Field Intelligence + KPI
Module C (4 forms) + Module D (KPI engine, deductions, incentive brackets, all dashboards, Google Sheets) · Pen test findings resolved
✓ Full KPI engine live
WEEKS 13–17 · STAGE 4
Crystal + Go-Live
Module E (Crystal queue, fake check detection, coaching log) · All integrations final · UAT 10 pilots · Training 60 amb + 2 TL + HSS · Go-live Riyadh + Jeddah
✓ September 2026 go-live
Client dependencies — required at kickoff to protect this timeline
RiderCare API specification and credentials  ·  WMS (Rooster) API access  ·  FMS GPS API credentials  ·  HungerStation IDP for SSO  ·  10 pilot ambassadors for field UAT in Stage 4