All customer storiesProduction system
99.25% of 4.6 million transactions a month, matched automatically.
A betting and gaming operator reconciled about 4.6 million transactions a month across several payment providers, largely by hand. AI now reads every provider file into one format, five rule-based passes match 99.25% automatically, and finance works only the exceptions.
- Client
- A betting and gaming operator
- Sector
- Gaming, finance operations
- Engagement
- Production system
- Measured result
- 99.25%match rate, 4.6M transactions
99.25%match rate, 4.6M transactions
The full account
01More providers, more formats
The operator was adding payment methods to reach players in more markets, and each provider sent a settlement file in its own format. About 4.6 million transactions a month were matched by hand in spreadsheets, and only a few people could read every file. Reconciliation lagged the business and was a control risk.
02AI for intake, rules for matching
AI does intake only, because that is where the input varies. It maps each provider file, irregular layouts and noisy references included, into one standard transaction record. Matching is a job for rules: five deterministic passes, each handling a defined class of match and passing the rest to the next. The rules can be read, tested and explained to an auditor, and change only through change control.
03What finance has now
The engine matches 99.25% of transactions automatically, each match linked to its source lines and the rule that made it. What remains goes to a named owner, with the passes already tried and the fields that disagree. Finance works the exceptions instead of matching routine transactions, and a new provider needs a new intake mapping, not new matching rules.