Artificial intelligence news.
Straight from the original announcements.

Resemble AI launches Agent Detection to tell humans, bots and AI agents apart

October 29, 2026By the AI News Bites Editorial TeamSource: Resemble AI
Resemble AI launches Agent Detection to tell humans, bots and AI agents apart
Image: Resemble AI

Resemble AI has introduced Agent Detection, a tool that analyzes visitor behavior and intent to classify every session as human, bot or agent, and explains its reasoning. The company says traditional bot detectors cannot separate a customer's AI agent from a malicious one, so both end up blocked.

Key points
  • Resemble AI's Agent Detection classifies each session as human, bot or agent, with reasoning.
  • The company says traditional bot detectors rely on metadata such as IP address and VPN and only return human or not human.
  • Permitted agents can be routed to the right place, while unauthorized agents can be kept from sensitive data and fraud.
  • Resemble says Meta's Muse reached 5 million downloads in under a month, versus nearly two months for ChatGPT.
  • A free agent scan is available through Resemble AI.

A new kind of visitor on the web

Resemble AI describes AI agents as "the fastest-growing visitors on the web". According to the company, they browse, log in and complete tasks on behalf of customers or bad actors. Earlier bots were built to read content or run simple, repetitive tasks. Agents can carry out a complex series of tasks, either for a person as a personal agent or autonomously and at scale, opening multiple website and product sessions in parallel.

Resemble points to the speed of adoption as a sign of where traffic is heading. In its video, the company says Meta's AI agent, Muse, saw 5 million downloads in under a month, while it took ChatGPT nearly two months to reach that number. Resemble adds that anyone can have an agent today, and that intent can be good or bad.

Why bot detectors fall short

According to Resemble, existing bot detectors rely on metadata such as IP address and VPN use. That information cannot distinguish a person from an agent running in a real browser. Their verdict is also limited to human or not human.

The company says browser-based agents slip past these tools easily because they perform multiple actions toward a single goal, rather than one action on repeat. Even when an agent is flagged, there is no way to tell a customer's agent from a malicious one, so both are blocked. Resemble also claims that most organizations do not know how many agents are on their site or in their product at any given time.

How Agent Detection works

Resemble already detects AI in content and in communications. With Agent Detection, it extends that to behavior. The system analyzes how each visitor moves and what their intent is, then labels the session as human, bot or agent. It provides reasoning with each classification so the site owner can decide how to handle it.

The company outlines two main uses:

  • Routing permitted agents to the right place to complete their task.
  • Preventing unauthorized agents from reaching sensitive information or completing fraudulent attempts.

Resemble argues that behavior is the newest modality that requires AI detection, alongside voice, image and video. It also notes that agents rarely stay on one site. Because they can log in on a user's behalf, make phone calls, send emails and upload files, the company says it offers detection models covering content, communications and behavior. Resemble states that, like all its models, the AI detection is built in-house.

What it means for site owners

For organizations that run websites and products, the pitch is visibility and control: knowing what is accessing their properties and treating helpful agents differently from harmful ones, rather than blocking everything that is not a person. Resemble is offering a free agent scan through its website so companies can see who is visiting. The source does not give pricing for paid use or details on availability beyond that scan.

Related news

For developers
For developersNovember 13, 2026, 18:00

Sierra releases Poppy 1.0 for personal agents and businesses

Sierra has released version 1.0 of Personal Agent Protocol, known as Poppy, giving personal agents a standard way to interact with companies through business agents…

Products and tools
Products and toolsNovember 12, 2026, 19:21

RemotePC Adds MCP Integration for AI-Assisted Computer Management

RemotePC has introduced an integration that lets customers use ChatGPT, Claude and other compatible AI assistants to monitor remote computers and perform supported…

Resemble AI launches RAD-S1 to spot AI agents that pose as humans
Safety and ethicsOctober 29, 2026

Resemble AI launches RAD-S1 to spot AI agents that pose as humans

Resemble AI has released RAD-S1, a behavioral model that tries to tell whether a website visitor is a person or an AI agent, even when the agent is disguised. According…