Skip to content

Fraud use cases for banks and fintechs

15 attacks, each told as it happens: how it unfolds, which checks it passes on its own, what DeepID reads, the decision it returns and the words the customer sees.

Scams

The genuine customer is deceived and authorises the payment herself. Every coached payment needs three things: a call to keep her on, a payee to send the money to, and checks that she passes because she is real.

  • Digital arrest

    Hold the coerced payment before it leaves the account.Hold, then call back
  • Investment scams

    Slow the scam payment and hold the mule trading account.Warn, then cool off
  • Vulnerable customers

    Extra time, and a call back, for the customers most at risk.Cool off, then hold

Mules

Stolen money needs accounts to land in and routes to leave by. Each account looks ordinary on its own; the ring shows when faces, devices and payees are linked.

Deepfakes

A generated face or a built identity passes a check made for photos. The evidence sits below the picture: the camera, the device and the form.

  • Video KYC deepfakes

    A live person, in front of a real camera, on every video KYC call.Step up, hold or block
  • Executive deepfakes

    Verify the approver on their own phone before an unusual payment leaves.Step up
  • Loan stacking

    Catch stacked and synthetic applications before disbursal.Cool off or block
  • Claims deepfakes

    Pay clean claims without an extra step, and hold the fabricated ones.Hold, or allow

Account access

The attack arrives through recovery, an agent, an employee or software, with valid credentials. The person holding them is the question.

  • Account takeover

    Protect the account through recovery and the hours after.Step up, then cool off
  • Assisted channels

    Prove the customer is there, on her own phone, before an assisted action.Step up, or hold
  • AI agents

    Tell the customer’s own assistant from automation that only looks like one.Allow, or step up
  • Insider fraud

    Hold the takeover that starts with a contact change made from the inside.Step up, then hold

One pattern runs through all 15.

Each check looks at one moment, and each moment looks clean. The attack shows only when the person, the phone and the payment are read together across the journey: a live person on her own phone, acting by her own choice, paying a payee away from mule rings.

That is why every page sets what your checks saw beside what DeepID read at the same step, and ends in one of six actions with its reasons, not a score.

How it works, moment by moment

Abuse on consumer apps.

Not scams on a bank account but abuse of the app itself: one person behind many accounts, offers claimed again and again, and bots copying what you publish. The device behind them is what links them.

  • Promo abuse

    Each offer claimed once per real customer.Step up, or block
  • Multi-accounting

    Many accounts on one phone, linked at sign-up.Warn, step up or block
  • Fake accounts

    Fake profiles stopped before they post or collect.Step up, or block
  • Referral fraud

    Referral loops stopped before they pay out.Step up, or hold
  • Account sharing

    Sharing seen, and loyal members left alone.Warn, or step up
  • Content scraping

    Catalogues and prices kept away from bots.Step up, or block

Start with what your history can show.

A free retrospective on 30 to 90 days of your own transactions shows the payee and mule links in them. Session signals follow in a shadow proof, once the SDK runs.