Case study · Sportsbook · West Africa · 2026
Premier Bet
Small operator, mobile-money discipline, four-language ops.
A 110-person Premier Bet subsidiary across two West African markets embedded AI in mobile-money reconciliation, retail-agent integrity, USSD flows, and multi-language support — without weakening the regulator-facing answer in either market.
At a glance
- 110Headcount
- 4 weeksEngagement
- 9Pioneers
- 2Markets
- 4Money rails
The brief
AI rollout where the rails are mobile money and the channel is USSD.
Cohort
- C-level (CEO, COO)2
- Heads of Department5
- AI pioneers9
Drawn from Trading, Player Ops, Payments, Retail, CRM, and Compliance — a small team running two markets, four mobile-money providers, and a USSD-first channel mix.
What the pioneers shipped
Six workflows tuned for mobile money and USSD-first players.
Mobile-money reconciliation
Reconciles deposits and withdrawals across M-Pesa, MTN Mobile Money, Airtel Money, and Orange Money. Finance reviews the exception queue, not every line.
Multi-language support copilot
Drafts replies in French, English, Wolof, and Lingala based on the channel the player came in on. Agents review tone and send.
USSD flow optimiser
Identifies drop-offs in the USSD bet-placement tree, drafts reroute candidates. Product team picks one, ships behind a flag.
Retail-agent integrity watch
Flags abnormal cash-in / cash-out patterns at retail kiosks before settlement. Compliance reviews; agent network team acts.
Per-market RG calibration
Calibrates harm thresholds per regulator — limits, cool-off prompts, self-exclusion routing differ between the two markets without leaking data between them.
Disputes & chargeback triage
Categorises mobile-money disputes by provider, drafts the response with cited transaction context — payments analyst signs off.
“Two regulators, four money rails, four languages. Speed without breaking the answer in any of them.”
Approach
Four weeks. Two regulators. Mobile-money rails untouched.
Discover
Mapped both market stacks, the four mobile-money integrations, the USSD tree, and the retail-agent settlement window. Compliance defined what an agent could draft and what stayed human in each market.
Build
Pioneers shipped first cuts inside their function. Anything that touched a player flow or a money rail paired with the relevant lead before staging — not after.
Per-market sign-off
Two demos with the local Heads of Compliance, one per regulator. Anything not aligned with local RG, KYC, or settlement rules was redesigned — not deployed.
Compound
Skills shipped to an internal marketplace. A win in one market lifts the other — without leaking PII or commercial terms between the two licensed entities.
Impact · Year one
Outcomes leadership signed off on.
Mobile-money recon
−82%
Daily exception queue cleared in 40 minutes instead of half a day.
Support first response
−66%
Across French, English, Wolof, and Lingala — same headcount.
Business satisfaction
9.2 / 10
Post-engagement leadership review, both market GMs.
Why it didn't go off the rails
Small team. Two regulators. Zero room for cute.
No agent action on money rails
Agents draft reconciliation, dispute responses, and refund proposals. A finance lead signs every transaction-touching outcome.
Per-market data isolation
No leakage between the two licensed entities. Each agent runs scoped to a single market or aggregates with PII stripped.
Language fidelity reviewed by humans
Drafted replies in local languages reviewed by a native speaker on the support team before send. Always.
Retail-agent integrity > automation speed
Anything that touches kiosk settlement goes through compliance and the agent-network team. No exceptions.
Where to start
Book an org assessment.
Ninety minutes, no slides.
We'll spend an hour with your leadership team mapping where AI creates the most leverage in your operation — and what an overlay would look like in practice.