LONDON, November 12th, 2026. Customers are starting to ask AI assistants to sign in to their bank, check balances, and make payments for them. To a bank, that traffic looks automated, and so does a credential-stuffing bot and an account aggregator pulling data for a budgeting app. Treat all three the same, and a bank either shuts out customers who chose an agent or lets fraud through.
Callsign has extended Digital DNA to separate them. When a web sign-in shows strong signs of automation, the Callsign Platform checks whether it comes from an account aggregator such as Plaid, Finicity, Yodlee, eMoney and MX from an AI agent acting for a customer, or from a scripted bot. The bank can then treat each one differently. The classification has been running on live banking traffic over recent months and is already identifying AI agents such as Instinct and Muse, among others.
Why banks need to see this traffic
One in five US adults (19%) say an AI agent has already asked them for login credentials or card numbers, according to a national survey conducted by Morning Consult on behalf of the American Bankers Association and released in October 2026. Some of those requests come from tools customers chose. Some come from scammers. Customers can’t always tell which is which, and neither can most banks. Some large US banks have responded by blocking AI agents outright. That stops the fraud, but it also blocks customers who chose to use one.
How it works
1. Score the session. The Callsign Web SDK reads device and behavioral signals as the customer signs in and produces a bot score.
2. Classify the automation. When the score is high, the platform checks for known account aggregators and looks for the signs of an AI agent. It labels the session aggregator, suspected AI agent, or bot, and shows the fraud team the reasons behind each label, so the decision is explainable. Agent and automated sessions never shape the customer’s own behavioral profile, so the Digital DNA that recognizes a person stays theirs.
3. Apply the bank’s own policy. Callsign raises the label, and the bank’s existing rules decide what happens. A typical policy allows approved aggregators, denies bots, and sends suspected agents to the customer for approval.
The customer approves their own agent
When a suspected agent tries to sign in, the bank can ask the customer to approve it in the mobile banking app they already use. The customer doesn’t get a code or a link by text, and nothing asks them to do the things banks have spent years warning them against. If the customer approves, the agent carries on. If they didn’t send it, it goes no further.
That approval is the agent’s lineage: the link back to the genuine person who sent it. An agent’s trust is only as strong as its lineage.
“Banks are being asked to choose between blocking every AI agent and trusting every one,” said Zia Hayat, CEO and founder of Callsign. “Neither works. A bank needs to know what is signing in and whether the real customer is behind it. We give the bank that answer and let the customer approve their own agent, so genuine users and their agents get on, and fraud gets stopped.”
“Detecting agents is only the start,” said Senan Moloney, Chief Product and Technology Officer at Callsign. “What matters is what the bank does next. We classify the automation, show the fraud team why, and when it’s an agent, we go back to the real customer on the device the bank already trusts. A bot gets stopped. An agent gets a chance to prove that the customer sent it.”
About the Callsign Platform
The Callsign Platform sits behind every Callsign solution.
It recognizes people by their Digital DNA and is extending that Digital DNA to identify AI agents acting on their behalf, tracing each agent’s lineage back to the genuine person who sent it.
That lineage is how we protect every transaction from fraud. An agent with a clear lineage will get through without friction. One with a missing or broken lineage will be challenged or stopped before money moves.
Know who’s real. Stop who’s not.
About Callsign
Callsign helps organizations recognize their users just like people do in the real world, naturally, securely, and without friction. Every person has a Digital DNA. A unique pattern they create through their devices, locations, and behavior. As customers start sending AI agents to shop, pay, and manage accounts for them, Callsign is extending Digital DNA to recognize those agents too, and to trace each one back to the genuine person who sent it. That’s how Callsign tells genuine customers from fraudsters in real time, whoever or whatever is behind a session, and protects every transaction from fraud and scams.
With Callsign Identify and Callsign Protect, organizations manage customer identity and access, and stop account takeover, scams, and fraud in real time. Genuine users and their agents get on with their digital lives, and organizations build trust into every interaction. Callsign is trusted by leading financial institutions and global enterprises. Trust begins with Digital DNA.
www.callsign.com
Media contact
For media enquiries, please contact Callsign's marketing team at peter.gabriel@callsign.com
