FingerprintJS Inc., a leader in device intelligence for fraud prevention, today announced the launch of a new ecosystem of signed artificial intelligence agents, enabling enterprises to distinguish between authorized agent AI traffic and malicious bots and scrapers.

“For years, the goal was simply to stop the bots, but that’s a losing strategy as an increasing number of interactions are becoming automated,” said co-founder and Chief Technology Officer Valentin Vasilyev.

For over a year, the terms AI agent and agentic AI have become constant words within the enterprise space, used to describe frameworks that allow large language models to take step-by-step reasoning and execute complex tasks. Now, with companies such as Google LLC also bringing these tools into browsers, consumers can also use these agents to allow their computers to shop for them.

Fingerprint argues that modern-day companies are facing an ever-increasing threat, although these types of bots are not currently designed to commit fraud. That may yet change in the future, in that they automate what humans do: Go out, look through listings, identify things to buy, put items in shopping carts and then execute purchases.

All of these automations mimic human behaviors on the web. Shopping is only one example of automation, other agents already automate research, stock trading, filling in job applications and monitoring social media — and make comments directly.

“We built this ecosystem so businesses can stop blindly blocking visitors,” added Vasilyev. “Instead, they can now start identifying every visitor, whether they are a malicious bot, an authorized agent or a human.”

Ideally, a business can allow through an automation that works for a person, watch one that is unknown for potential suspicious behavior (and block it if it misbehaves) and automatically block those already proven to have hostile signatures.

The company has already signed on ecosystem partners with OpenAI Group PBC, Amazon Web Services Inc.’s AgentCore, Browserbase Inc., and Manus, which was acquired by Meta Platforms Inc. in 2025. The ecosystem allows enterprises to detect authorized AI traffic with complete certainty.

“Open internet protocols mature through real-world deployment,” said Thibault Meunier, research engineer at Cloudflare Inc. and creator of Web Bot Auth. “Operating them at scale is how we validate assumptions, surface edge cases and improve the standard.”

AI agents built on platforms such as Manus can access permissioned environments, such as PitchBook Data Inc., to analyze data without being flagged as malicious immediately. It also allows for identity control because AI agents have their own identity and management control.

For example, e-commerce and fintech leaders can now selectively permit transactions for agentic buyers, while maintaining audit trails and real-time anomaly detection of payment fraud.

The company said it’s platform already identifies more than 1 billion unique devices every month and processes hundreds of signals to help fraud teams distinguish trusted visitors from bad actors with speed at scale. More than 6,000 companies, including Dropbox Inc., Checkout Ltd. and Neuro-ID Inc. use the service to identity high-risk activity in real time to prevent fraudulent activity and still deliver good experiences to users.

“This collaboration makes ‘agent identity’ a first-class concept on the web, so companies can confidently permit trusted agents while blocking the rest,” said Browserbase founder and CEO Paul Klein.

Image: Pixabay