Plaid launches Agent Detection to identify AI agent activity

18 November 2026
News
Plaid launches Agent Detection to identify AI agent activity

Plaid has launched Agent Detection, a tool that helps financial firms determine whether an interaction comes from a person or an AI agent. The product is aimed at fintechs, banks, and other financial companies. In addition to identifying the type of interaction, it also shows which AI provider the agent is associated with.

Plaid describes the launch as the first step in a broader effort to establish trusted agent interactions across the financial sector.

AI agents are entering financial channels

The launch follows the arrival, within a short period, of several consumer-facing AI agents, including Grok Bot, Meta Muse, OpenAI dots, and Claude Duet. They can operate in a browser and complete tasks on a user’s behalf.

According to Pedro Sanzovo, Head of Fraud at Plaid, agents are already appearing on login pages and in onboarding and account-linking flows. This makes it harder for companies to determine who is actually taking action during a session.

How Agent Detection works

AI agents can navigate websites in ways that resemble human behaviour. As a result, standard device and behaviour checks may not be enough to tell them apart. Agent Detection analyses thousands of signals related to devices, behaviour, automation patterns, and network activity to classify each interaction.

The model has been tailored to financial use cases. Plaid already analyses device, behaviour, and network data in financial login flows across hundreds of institutions. The company says this helps it distinguish an agent logging in to a customer’s bank account from one carrying out a shopping task.

In side-by-side tests on live traffic, Agent Detection identified 2.3 times more agent sessions than competing solutions, with no increase in false positives.

For example, a consumer could ask an agent to pay an upcoming bill. The financial institution could then recognise that an agent is behind the interaction and decide how to handle the payment initiated in this way.

Access to data with consumer permission

Plaid already provides consumer-permissioned access to financial data for thousands of apps, including AI assistants such as ChatGPT, Claude, Perplexity, Grok, Meta Muse, and Google Gemini. Users authorise access without giving the app their bank login credentials and can see what data is shared and why.

Access is read-only, uses existing bank-approved APIs, and is subject to the same data-use rules that apply to all Plaid customers.

Agents that act on a consumer’s behalf, rather than only retrieving information, raise additional questions about transparency, authorisation, and control. Plaid plans to work with financial institutions and agent developers to extend consumer-permissioned access into an approved way for AI agents to act. Security, visibility, and user permission will be central to this approach.

Two approaches to handling agent traffic

Plaid expects financial firms to take different approaches to supporting AI agents and is developing capabilities for both. For firms that want a dedicated channel for agents, the company is building on its consumer-permissioned access infrastructure to create a mechanism that lets consumers approve specific actions.

The design aims to give consumers visibility and control over the actions they authorise, while giving financial institutions greater confidence in the interactions they accept.

For firms already seeing agents interact with their existing digital channels, Plaid is expanding its Trust Index. The tool currently scores sessions for risk and is being developed to assess agent activity in context.

Using intelligence from across its network, Plaid aims to help firms distinguish legitimate agent activity from potential risks and decide how to handle each interaction.

Top