How do you cut fraud false positives without missing fraud?

At Payrix, false positives fell from 576 a day to 86 while detection held at 4,771 of 4,800, above the payment network floor of 4,752. The gain came from replacing twelve fixed rules with a model reading 140 features, not from loosening thresholds.

What does a high false positive rate cost?

$2.9M a year here in review labour, with 24 analysts working 4,800 daily reviews when around 14 transactions a day were actually fraudulent. Reducing reviews to 720 took that to $440K, and legitimate customers stopped being held up.

Why do fraud analysts need an explanation with each alert?

Because an alert with no reason attached starts the investigation from nothing. Each of the 720 daily reviews at Payrix now carries the top three fraud indicators, which took review time from 8 minutes to 3.

Payrix (Worldpay) · Financial Services · Austin, TX · 14 weeks

Fraud detection cut false positives from 576 to 86 a day

A B2B payments fintech flagged 576 out of 4,800 daily transactions as fraudulent, forcing manual review of every transaction. We rebuilt their fraud detection model, cutting false positives to 86 per day. This saved $2.9 million annually in review costs while still catching 4,771 of 4,800 actual fraudulent transactions.

Challenge

The rules-based fraud system flagged 576 out of 4,800 daily transactions, forcing 24 analysts to manually review every alert. This generated a massive false positive problem, only 14 transactions were truly fraudulent each day. Legitimate customers faced unnecessary friction, and the company wasted $2.9 million annually on analyst labor handling mostly false alarms.

Approach

We replaced the outdated rules-based system with a gradient boosting model using 18 months of transaction data. The model analyzed 140 features, such as transaction frequency, merchant categories, counterparty networks, and time-of-day trends. We added an explainability layer that clearly showed the top three fraud indicators for every flagged transaction.

Outcome

False positives dropped from 576 to 86 daily, cutting manual reviews from 4,800 to 720 transactions each day. This reduced annual review costs from $2.9 million to $440,000. We caught 4,771 out of 4,800 fraud cases daily, exceeding the payment network’s required threshold.

140 Features Our Rules-Based System Completely Overlooked

Alright, here’s the deal. The old system was running on just 12 fixed rules, pretty rigid, right? Not much wiggle room to catch the sneaky stuff. So, we flipped the script. We built a gradient boosting model with 140 features. Yeah, 140. We looked at things like how fast transactions move, where the counterparty sits in the network, and even the order of merchant categories. That gave us a whole new lens to spot patterns the old rules completely missed.

How We Made Review Decisions Clearer, and Kept Ourselves Accountable

We came across 720 transactions that were kinda iffy, could’ve been fraud, could’ve been nothing. Instead of digging through all of them one by one, we built a quick tool that flagged the top three signs of fraud for each case. That small move cut our review time from 8 minutes to just 3.

How We Nailed Accuracy on 4,752 of 4,800 Records Using Simple Data Checks

Alright, here’s what went down. The payment network wanted us to catch at least 4,752 fraud cases out of 4,800. No easy feat. So, we built a model that nailed it and then some. It flagged 4,771 fraud cases and kept false alarms super low, just 86 a day. Not too shabby, right?

Results

$2.9M Annual review labor saved
576 to 86 per day False positive rate
4,800 to 720 Daily manual reviews
4,771 of 4,800 Fraud transactions caught daily

We had 24 analysts reviewing 4,800 transactions every day, and 4,680 of those were legitimate. Thinklytics cut our false positives from 576 to 86 daily, while still catching over 4,752 fraudulent transactions. The ROI was clear from the start.
Chief Risk Officer, B2B Payments Fintech

In the client's terms

The question

Which of this work should a person still be doing?

The situation

A rules-based system flagged 576 of 4,800 daily transactions and 24 analysts reviewed every alert. Around 14 a day were actually fraudulent. Legitimate customers were being held up and the review bill ran to $2.9M a year.

What we found

The rules were not wrong, they were blunt. Each one had been added in response to a real incident and none had ever been removed, so the system was optimised for catching everything and nobody had priced the cost of catching everything.

What it produced

False positives fell from 576 a day to 86 and manual reviews from 4,800 to 720, taking annual review cost from $2.9M to $440K. Detection held at 4,771 of 4,800, above the payment network threshold.

The roadmap

A gradient boosting model trained on 18 months of transaction history across 140 features covering transaction frequency, merchant category, counterparty network and time-of-day pattern. An explainability layer returns the top three indicators per flagged transaction, because an analyst reviewing an alert with no reason attached is back where they started.

The work behind this result

Our systems need people in between them, the problem this engagement resolved.
AI Workflow Automation Consulting
AI Agent Consulting for Business Workflows
AI Automation Consulting
CRM Data Access and Rollback Checklist, for whoever has to approve the spend.
The 30 day Corporate Drag and Risk Diagnostic, findings yours either way.