Browser agents, desktop assistants and computer-use tools now open real browsers, pace themselves and reason through underwriting questions. ForMotiv tells you in session whether a person, a person with an AI assistant, or an autonomous agent is completing the application, and then what that agent is trying to do.
100+web-capable AI agents tracked
Dedicated R&Dupdating hundreds of features as new agents ship
Millisecondsto classify, before the policy binds
Why bot defenses miss them
Agents run in real browsers on clean devices, slow down to human speed and solve CAPTCHAs. What they can’t borrow is the way a person reads, aims, pauses and corrects themselves.
| Control | Scripted bots | AI agents |
|---|---|---|
| Headless-browser checks | Usually caught | Missed: real browsers |
| Speed and velocity rules | Usually caught | Missed: human-paced |
| CAPTCHA | Mostly caught | Solved |
| Device fingerprinting | Often caught | Rotated per session |
| User-agent strings | Sometimes | Spoofed or generic |
| Behavioral intelligence | Caught | Caught |
Explainability
Pick a source and replay it. The cursor trail and keystroke strip show what ForMotiv sees; the panel shows which families of behavior separate that session from a human one.
Agentic R&D
ForMotiv runs a dedicated research group and a controlled lab environment focused solely on agentic detection. Every time a new agent or model version ships, it goes through real insurance application flows, the same way it would hit yours.
Real flowsQuote, registration and servicing journeys like the ones carriers ship
Every surfaceBrowser, desktop, command-line and computer-use agents
Hundreds of featuresMouse, keyboard, timing, device and network behavior, refreshed continuously
Network effectNew patterns roll out to every carrier on ForMotiv
The lab picks it up as soon as it’s publicly available.
Quotes, logins and policy changes, repeated across many sessions.
Against human sessions at scale, feature by feature.
New or retuned behavioral features ship to carriers, with no change to your integration.
Repeats with every release
After detection
Plenty of AI traffic is good business: a real customer asking an assistant to finish a form. ForMotiv answers two separate questions, who is in the session and what they’re doing, so you never have to treat all AI as fraud.
| Who is in the session? | Quotes once, then buys | Requotes every day | Cycles answers | Harvests prefill | Stolen or synthetic identity |
|---|
Who it is and what it’s doing are scored independently, then combined under your own rules by line of business and channel. Configure the response once; ForMotiv applies it to every session in milliseconds.
Agents and agencies using AI
Producers use AI to quote more carriers and key applications faster. ForMotiv separates productive automation from harmful automation at the agent and agency level, by connecting how business is written to how it performs.
Volume up, conversion up, retention and loss experience unchanged. Reward it.
Volume up sharply, rewalks and answer edits up, policy quality falling. Review it.
ForMotiv Console
Fraud, product and underwriting teams track agentic sessions offline in the Console, then drill into any one of them.
Daily trend of sessions identified as agentic, by state and date range
Quote, bind and send-to-agent rates for that traffic
Each session tagged with the suspected tool family
Sessions grouped by operator to expose scaled activity
Related
The same classification feeds every channel and transaction, and joins the same behavioral record as every other session.
Personal AI as a new channel, and how carriers serve, shape or stop it.
Steering and prompt-injection attempts inside AI and live chat.
Bots, rings, ghost brokers and account takeover across the lifecycle.
Requotes, answer cycling and prefill harvesting where they start.
Credential testing and automated account creation on portals.
Pull the agentic features straight into your own rules and models.
What is agentic AI detection?
Agentic AI detection identifies when an AI agent, rather than a person, is filling out an insurance application, logging in or chatting, and reads what that agent is trying to do.
How does ForMotiv detect AI agents?
ForMotiv analyzes hundreds of constantly updated behavioral features to recognize when an AI agent is filling in a form or chatting, including agents that try to imitate human behavior. A dedicated research team keeps those profiles current as new agents are released.
Which AI agents can ForMotiv identify?
ForMotiv is tracking 100+ web-capable AI agents. It profiles browser agents, computer-use agents, headless automation and agent-assisted applicants, and adds new agents and versions as they appear.
Is detecting an AI agent enough?
No. Detection is only the start. What matters is intent: whether the agent is acting for a legitimate customer or abusing the quote flow, and what the carrier should do about it.
What should a carrier do when an AI agent applies?
Carriers can allow the session with guardrails, route it to a person, suppress costly data calls or block it, based on the agent's behavior and the carrier's own policy.
Does ForMotiv explain how it detects agents?
ForMotiv returns clear results and reasons for each session, but does not publish its detection methods, which helps keep them effective as agents evolve.
Bring a quote flow. We’ll show you the agentic sessions already in it, what they were trying to do, and the response we’d recommend for each.
Leakage, fraud, conversion or agent oversight, on the line and channel you care about most.
A walkthrough of the signals, features and scores for journeys like yours.
Where the tag goes, what the API returns and which systems act on it.
