AI Agent Authorization
LiveAn AI agent does not evade your checks. It passes them.
Every fraud signal ever built is a proxy for "is a human here", and an AI agent genuinely has the things those proxies measure: it drives a real browser, so the fingerprint is real; it arrives on a residential connection, so the address is clean; it controls a real mailbox, so it receives the code. Detecting it is an arms race nobody wins. So this asks a different question. Agents that identify themselves sign their requests with a key their operator publishes, and the operator's Signature Agent Card declares whether the request was initiated by a person or by autonomous scanning. That distinction — fetcher versus crawler — is invisible to every classical bot signal, and it is the difference between your customer's assistant doing what they asked and something acting on nobody's behalf.
What it detects
- Web Bot Auth signatures verified against the operator's published keys
- User-initiated fetcher vs autonomous crawler
- Allow, deny or rate-limit per operator, path and method
- Proof rather than inference — no false-positive rate
- An unverifiable signature treated as worse than none
When to use it
At any endpoint automation reaches — signup, checkout, search, content — before you decide whether to serve it.
Questions
Can you detect an AI agent that does not identify itself?
Not reliably, and anyone telling you otherwise is selling an arms race. A capable agent driving a real browser on a home connection looks exactly like a person because in every measurable respect it is behaving like one. What we can do is tell you with certainty when an agent DOES identify itself — and increasingly they do, because operators want their traffic accepted. Unsigned automation then stands out by omission rather than by detection.
Should I just block every AI agent?
That reflex gets more expensive every month. A growing share of legitimate traffic is somebody's assistant acting on their explicit instruction — booking, comparing, filling in the form they asked it to fill in. Blocking that refuses a customer who is standing right there. The useful line is not human versus machine, it is whether a person asked for this request.
Is this a standard or your own invention?
A standard. Web Bot Auth builds on RFC 9421 HTTP Message Signatures, with an IETF working group chartered in early 2026 and Cloudflare, Google, Amazon and OpenAI behind it. We implement it — we did not invent it, and we would rather you could switch away from us than depend on something only we speak.
What if an agent lies about being user-initiated?
Then it lies under a signature bound to its operator's identity, which is a reputational risk no serious operator wants and a fact you can act on. But the honest answer is that a truthful fetcher can still be abusive: someone really can instruct their assistant to open fifty trial accounts. The card cannot see that and does not claim to. The linkage signals on the unified score can, which is why we recommend using both.
What happens when an agent publishes no card?
You get review, not a guess. We know who they are cryptographically and nothing about what they intend, and saying so is more useful than inventing a verdict. Most operators have not adopted the registry draft yet, so this is currently the common case.