Case study · Sportsbook · 2025
BetForge
Sportsbook trading desk goes AI-native, without slowing the lines.
A multi-market sportsbook with 400 staff embedded AI across odds compilation, risk, and support — same risk veto, faster cadence.
At a glance
- 400Headcount
- 4 weeksEngagement
- 10Pioneers
- 6Markets
- 24/7Operating
The brief
AI inside the trading desk, with risk veto on every line.
Cohort
- C-level (CEO, COO)2
- Heads of Department5
- AI pioneers10
Drawn from Trading, Risk, Payments, Customer Support, and BI — the five functions where speed and accuracy compound.
What the pioneers shipped
Bots that think with traders, not for them.
Live trading copilot
Watches markets, suggests line moves with rationale. Trader signs every change.
Risk-team alert triage
Surfaces unusual betting patterns, links to similar past cases, drafts the analyst note.
Settlement reconciliation
Cross-check between trading platform and payments rail. Discrepancies flagged before payout.
Customer support classification
Tier-1 questions auto-routed; complex disputes escalated with full context.
In-play exposure summary
Natural-language summaries of in-play volumes and book exposure — refreshed every 60 seconds.
“Faster cadence — same risk veto on every line.”
Approach
Four weeks. Risk team in the room from week one.
Discover
Mapped trading workflow, risk gates, settlement rails. Risk team and traders co-defined what could be assisted vs. authored.
Build
Pioneers built first cuts inside their function. Every model output passes through a risk-team review before staging.
Validate
Three Friday demos. Each model challenged by traders running their own A/B against the agent's suggestions.
Compound
Skills shipped to internal marketplace. Cross-market learning loop instrumented; suspended-market actions stay fully manual.
Impact · Year one
Outcomes leadership signed off on.
Annual savings
€3.2M
Across the engagement's first twelve months.
FTE optimized
24
Full-time equivalents freed for higher-leverage work.
Business satisfaction
9.4 / 10
Post-engagement leadership review.
Why it didn't go off the rails
Trader veto. Risk veto. No exceptions.
Traders sign every line move
Agent proposes; trader signs. No automated price changes on live markets.
Risk team has model veto
Each model has a risk owner. Anything not signed off by risk doesn't ship.
No automation on suspended markets
Suspended markets are always handled manually. Audit trail per intervention.
Cross-jurisdiction logging
Same query in different markets returns the same documented answer — every regulator inspection passes.
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.