ATO

Account takeover fraud: the complete guide to prevention, detection, and response

Published
November 16, 2026
Read Time
12
mins
Gal Perelman
Gal Perelman
Product Marketing Lead, Unit21
Subscribe to stay informed
Table of contents

The first sign of an account takeover is often a phone call. A customer can't log in, or they're asking about a transfer they never made, and by the time the call reaches your fraud team the money has already left. Account takeover fraud is hard to get ahead of because the attacker is using a real account that belongs to a real customer, often with the right password.

‍

This guide covers how account takeover attacks work, the warning signs worth watching, and the controls that prevent account takeover fraud at banks, credit unions and fintechs. It also covers what happens once an attack is underway: how to detect it in real time and how to investigate the alerts that follow.

‍

‍

What is account takeover fraud?

Account takeover (ATO) fraud is when a criminal gains control of a legitimate customer's account and uses it for their own ends. Usually that means moving money out, opening new credit, or changing the account's details so the real owner is locked out. Most takeovers start with stolen or guessed credentials. Some start with a customer who has been talked into handing over a one-time passcode.

‍

ATO is a form of identity theft, but a narrower one. Identity theft covers any misuse of someone's personal information, including opening new accounts in their name. Account takeover means hijacking an account that already exists, which is part of what makes it dangerous: the account has history, trust and often a balance. Our dictionary entry on account takeover fraud covers the definition and well-known examples.

‍

Why account takeover is getting harder to stop

Credentials are cheap. Years of data breaches and password reuse mean many customers' logins are already circulating. Attackers test them at scale with automated credential stuffing, and only need a small share to work.

‍

The signals can be faked. Proxies, emulators, device spoofing and location spoofing let a fraudster look like a returning customer logging in from the right city.

‍

Money moves faster. Instant payment rails shrink the gap between takeover and loss to minutes. Once a real-time transfer settles, getting it back is difficult.

‍

Social engineering scales. Phishing kits and AI-written messages make convincing lures cheap to produce. Some attacks skip the password entirely by persuading the customer to approve a login or read out a code.

‍

The result is that no single control holds on its own. A strong password can be phished, and a one-time code can be intercepted. Prevention has to be layered.

‍

How account takeover attacks work

Most takeovers follow the same three stages, whatever the entry point:

  1. Access. The attacker gets credentials, or a way around them.
  2. Control. They change the email, phone number, password or 2FA settings so the real customer can't get back in or won't be warned.
  3. Cash out. They move money, add a payee, request a new card, or use the account as a mule to pass other stolen funds through.

‍

The entry point varies:

MethodHow it worksWhat it exploits
Credential stuffingAutomated tools try leaked username and password pairs across many sitesCustomers who reuse passwords
Phishing and social engineeringFake login pages, texts or calls trick customers into sharing credentials or one-time codesTrust, urgency, and increasingly convincing AI-written lures
SIM swapThe attacker convinces a mobile carrier to move the victim's number to a SIM they controlSMS-based one-time passcodes
Malware and infostealersSoftware on the customer's device harvests credentials or session tokens, and can hide security alerts from the customerCompromised phones and computers
Support channel abuseThe attacker uses personal data from public sources to pass knowledge-based questions and reset accessCall centers and manual recovery flows
Remote access scamsThe customer is persuaded to install a remote-control app, and the fraudster drives the session directlyA session that looks legitimate because it runs on the customer's own device

‍

Former fraudster Alexander Hall describes how these methods combine in practice in our guest post on how account hijacking unfolds. For the device side, including emulators and bots, see our post on real-time device intelligence.

‍

Warning signs of account takeover

Individual red flags are rarely conclusive. Customers buy new phones, reset passwords and make large transfers all the time. What separates a takeover is the combination of signals and the order they appear in. They tend to cluster in three places:

  • The session. A newly seen device, a time zone or location that doesn't match the IP address, location spoofing, emulators, bot activity, remote-access tools, or anonymizing networks like Tor.
  • Account changes. Email, phone, password or 2FA changes, often within minutes of each other, or a burst of failed logins followed by a success.
  • Money movement. A first-time payee, a transfer or withdrawal soon after account changes, or a dormant account that suddenly empties.

‍

For more on individual signals, see the biggest risk signals for account takeover fraud.

‍

How to prevent account takeover fraud

Prevention works in layers, and each layer covers a gap in the others. These are the controls that matter most for banks, credit unions and fintechs.

‍

Use risk-based authentication

Every account should have multi-factor authentication. It stops a large share of credential-based attacks. But not all factors are equal. SMS codes can be intercepted through SIM swaps or phished in real time, while authenticator apps, passkeys and device binding are harder to steal.

‍

Blanket friction has a cost too. If every login triggers a challenge, good customers get annoyed and some leave. The better pattern is risk-based: let a returning customer on a known device through quietly, and ask for step-up authentication when the session looks wrong.

‍

Protect the moments when accounts change

In a takeover, the login is only the start. The changes that follow are where it becomes permanent: a new email, phone number, password, 2FA method or payee. Put checks at exactly those points. Confirm changes through a channel the customer already trusted, alert on changes made from a new device, and consider requiring step-up authentication on the first payment to a newly added payee. Fraud leaders we spoke to for the Fraud Fighters Manual make the same point in our roundup of account takeover prevention tips.

‍

Score the device and the session

Device intelligence identifies the device behind a session and flags tampering, emulation, remote control or spoofed location. Those signals matter most when they become fields your rules can act on. Unit21's device intelligence comes from the Unit21 Device SDK and produces a device risk score you can use directly in detection rules.

‍

Put real-time controls on money movement

Because cash-out happens fast, some controls have to sit in the payment path itself. A real-time rule can deny a transfer, require step-up authentication, or allow it and open an alert for review. Rule evaluation in Unit21's real-time payment fraud prevention engine runs in under 150ms (as of September 2026), so several checks can run before a payment decision is due.

‍

Use intelligence from beyond your own data

A fraudster who is new to you may not be new to the industry. Consortium data flags customers, devices or accounts that other institutions have already seen in fraud, which gives you a warning on the first attempt rather than the third.

‍

Educate customers, but don't depend on them

Customers who use unique passwords, a password manager and stronger authentication are much harder to take over. Tell them plainly that your staff will never ask for a one-time code, and keep teaching them to spot phishing. But education is one layer, not the plan. Malware can hide your security alerts from the customer entirely, and social engineering is built to beat careful people.

‍

‍

How to detect account takeover in real time

Prevention won't stop every attempt, so detection has to catch what gets through, ideally before the money leaves. Effective ATO detection combines session and behavior signals in sequence, runs in the payment path, and gets tested before it touches live customers. Fraud teams can now describe a rule in plain English, validate it against historical data, and run it in shadow mode before it acts on anyone.

‍

We walk through the full approach, including the live build from our masterclass, in our guide to account takeover fraud detection.

‍

How to investigate an account takeover alert

When an ATO alert fires, the analyst has one question to answer: is this still the account holder? A practical checklist:

  1. Rebuild the timeline. Put logins, device and IP changes, and credential and contact changes in order.
  2. Compare before and after. Did transaction patterns change after the access change? Look for new payees, new payment rails and larger amounts.
  3. Check the device and the network. Is the device linked to other accounts or other alerts? Are several accounts paying out to the same place?
  4. Reach the customer through a trusted channel. Use contact details from before the change, never the ones that just changed.
  5. Act and record. Block, hold or release, and document the reasoning for audit and for any SAR decision.

‍

Much of steps one to three is assembly work. Unit21's AI investigation agents pull the history, run checks like counterparty analysis and behavior change, and draft a narrative, so the analyst starts with the evidence and spends their time on the decision. Our AI Task Spotlight on account takeover shows a before-and-after comparison task built and backtested on real alerts.

‍

Account takeover in specific situations

Dormant accounts

A dormant account that wakes up with a credential change and an outbound transfer is a classic takeover pattern. Nobody is watching the account, so the customer is unlikely to notice quickly. We looked at reactivation data in our post on account dormancy and fraud detection.

‍

Crypto and instant payment rails

On crypto platforms and instant rails, there is often no way to pull funds back once they leave. Prevention and real-time decisions carry more of the load. See how Gate US built its compliance program around real-time detection.

‍

How to choose an account takeover fraud solution

If you're evaluating tools, these questions separate a platform that holds up from one that adds noise:

  • Can it combine device, login and transaction signals in the same rule?
  • Does it make decisions in the payment path, fast enough for instant rails?
  • Can your fraud team build and change rules without an engineering ticket, and test them against historical data first?
  • How does it keep false positives down, so good customers aren't blocked?
  • What happens after the alert? Does it help investigate, and can you see exactly how it reached its conclusion?
  • Does it give you intelligence beyond your own data?

‍

Unit21 brings these together on one platform: device intelligence, real-time rules, AI agents that do the first pass on investigations, and AI rule recommendations that tune rules from analyst outcomes. Customers have seen up to 60% fewer false positives from rule recommendations (as of March 2026), and analysts stay in control of every decision. See how Unit21 approaches fraud prevention for banks and fintechs.

‍

Frequently asked questions

Is multi-factor authentication enough to stop account takeover?

No, although it stops many attacks. SMS codes can be intercepted through SIM swaps or phished in real time, and some scams persuade the customer to approve the login themselves. MFA works best alongside device signals, protected account changes and real-time controls on money movement.

‍

How do you prevent account takeover without adding friction for good customers?

Make friction risk-based. Let known customers on known devices through, and step up authentication only when several signals point the same way. Test new rules in shadow mode first, so you can see how many real customers a rule would stop before it stops any.

‍

What is the difference between account takeover and identity theft?

Identity theft is the broad misuse of someone's personal information, including opening new accounts in their name. Account takeover is a specific form of it, where the criminal takes control of an account that already exists.

‍

How do fintechs prevent account takeover?

With the same layers as banks, adapted to a digital-only model. Fintechs lean heavily on in-app authentication, device intelligence and real-time payment controls, because there is no branch to fall back on and many run on instant rails.

‍

What should you do when an account takeover is confirmed?

Lock the account and end active sessions, stop or recall pending payments where possible, and restore the customer's access through a verified channel. Then review linked accounts and devices, and document the case, including whether a SAR is required.

‍

Building an account takeover program that holds up

Account takeover rarely beats every control at once. It finds the gap: the reset flow without a check, the payee change nobody was alerted to, the rule nobody had time to tune. Layer prevention around the moments that matter, detect the sequence in real time, and give analysts the evidence to decide quickly.

‍

To see it built live, watch our account takeover masterclass. To see how it would work on your own data, get a demo with our team.

‍

Gal Perelman
Gal Perelman
Product Marketing Lead, Unit21

Gal Perelman is the Product Marketing Lead at Unit21, where she spearheads go-to-market strategies for AI-driven risk and compliance solutions. With over a decade of experience in the fintech and fraud sectors, she has led high-impact launches for products like Watchlist Screening and AI Rule Recommendations.

Previously, Gal held marketing leadership roles at Design Pickle, Sightfull, and Lusha. She holds a Master’s degree from American University and a Bachelor’s from UCLA, and is dedicated to helping banks and fintechs navigate complex regulatory landscapes through innovative technology.

‍

Learn more about Unit21
Unit21 is the leader in AI Risk Infrastructure, trusted by over 200 customers across 90 countries, including Sallie Mae, Chime, Intuit, and Green Dot. Our platform unifies fraud and AML with agentic AI that executes investigations end-to-end—gathering evidence, drafting narratives, and filing reports—so teams can scale safely without expanding headcount.
Risk and Compliance Operations
|
8
min

Build, buy, or both: how Circle splits its fincrime AI stack

Gal Perelman
Gal Perelman
Product Marketing Lead, Unit21
This is some text inside of a div block.
Risk and Compliance Operations
|
8
min

What examiners will ask about your AI now that model risk guidance leaves it out

Gal Perelman
Gal Perelman
Product Marketing Lead, Unit21
This is some text inside of a div block.
Risk and Compliance Operations
|
8
min

Running AI agents in production: how Rippling and Zepz prove quality every day

Gal Perelman
Gal Perelman
Product Marketing Lead, Unit21
This is some text inside of a div block.
See Us In Action

Boost fraud prevention & AML compliance

Fraud can’t be guesswork. Invest in a platform that puts you back in control.
Get a Demo