Fighting Fraud in a Connected World: paper checks, online banking, and shared intelligence. Presented by Matt Conine, founder of PayLynxs, at the Credit Union Association of New Mexico on November 3, 2026.
Check fraud hasn’t faded with check volume, and online banking has opened a new kind of scam. Matt’s session looked at both, and at why credit unions catch more fraud when they share what they see.
Checks are declining. Check risk is not.
- In the Federal Reserve’s 2026 Risk Officer Report (a Q4 2025 survey of 400+ financial-institution risk professionals), 63% said their institution saw check-fraud attempts in the prior 12 months. Respondents most often saw rises in counterfeit checks (32%), check washing (21%) and payee forgery (18%).
- Americans still wrote 9.2 billion checks in 2024, worth $24.45 trillion, or $2,653 each on average (Federal Reserve Payments Study, July 2026).
- One stolen check, from a Mitek case study: a mailbox broken into on January 22, 2025; the check for sale on Telegram for $250 on January 24; $11,300 in altered checks deposited by phone on January 29; then $64,500 in counterfeits presented through clearing on January 30, before the account was closed.
Convincing fakes are now easy to make
Matt showed what a general AI tool did when asked for a realistic sample check. It declined, then offered a fictional training check marked “non-negotiable,” and produced one in about 30 seconds. A stolen check also gives away a name, address, signature, and account and routing numbers, which feed impersonation, payment redirection and account takeover. The takeaway: don’t rely on “it looks real” as a primary control.
Three ways to see check fraud
- Account-based intelligence looks at account status and history.
- Single-institution image analysis looks at an institution’s own checks.
- Consortium image intelligence compares check images with what many institutions have seen. PayLynxs partners with Mitek, whose Check Fraud Defender works this way.
Matt’s point: it’s a question of visibility, not a feature list. No single control wins. Strong programs layer member controls like Positive Pay, front-line review, image and transaction analytics, case history, and intelligence shared beyond the institution.
Online banking is the MVP for scams
In a scam, the member is real: familiar device, correct password, normal sign-in. The scammer decides who gets paid, how much and when. The aim is to judge risk in real time: hold the riskiest payments, send suspicious ones to an investigator, and let routine activity through. Three questions for your digital banking vendor:
- What monitoring do you offer for online sessions?
- Do you offer, or plan to offer, real-time interdiction?
- Which integrations help with online and mobile fraud detection?
A network can see what one institution can’t
The session used a fictional exercise: two credit unions each working a fraud case, with contact details that only connect when the cases are shared. Nobody had bad fraud detection. Nobody had enough data. Sharing adds context, speed and pattern recognition, so a warning can reach the second or third institution before it learns the same lesson the hard way.
PayLynxs’s Peer Fraud Network launched in 2026. As of the November 3 session, Matt reported 67 participating institutions, 4,536 fraud cases shared, about 1,326 active investigations at any time, and 22 matches across institutions.
Three ideas to take home
- Checks are declining. Check risk is not.
- Online banking is the MVP for scams.
- A network can see what one institution can’t.
Matt will give a similar session at the Hawaii Credit Union League 2027 Annual Convention in April 2027.
