AI-powered detection
AI analysis runs behavioural analysis per visit and classifies visitors as real, crawler, bot, competitor or click farm. It complements score-based blocking in ambiguous cases and can raise the threat score for a person already known to be bad. Enterprise plans add async in-depth reports and a natural-language "Ask" interface.
Score-based detection is fast and deterministic, which is what you want for the clear-cut cases. AI analysis exists for the rest — the visits where the signals conflict and a rule would guess wrong either way.
What it does
- Runs behavioural analysis on a visit and classifies the visitor into one of the five traffic categories.
- Optionally triggers blocks on its own verdict, so genuinely ambiguous traffic still gets acted on.
- Boosts the threat score for a person already established as bad, so their later sessions are caught faster.
Enable it per website in Settings → Protection. It complements click-based blocking rather than replacing it — both remain active.
Enterprise: reports and Ask
- In-depth AI fraud reports
- Generated asynchronously across a period of traffic, surfacing patterns that are invisible in any single session.
- Ask
- A natural-language interface for questions about your traffic — ask what you would ask an analyst, in plain English.
Availability
AI analysis is available on paid plans; in-depth reports and Ask are Enterprise features. See Plans and limits.
Frequently asked questions
Does AI detection replace the threat score?
No. The score still drives the primary blocking decision; AI analysis adds a second opinion for ambiguous traffic and strengthens scoring for known-bad people.
Is my visitor data used to train models?
Analysis is performed to classify your traffic. Credentials and tokens are encrypted at rest and never exposed to the browser; see ConversionOS and privacy for how identifiers are handled.
Related articles
How detection works
Person-based fraud detection: fingerprinting, identity clustering across IPs and devices, and the 150+ behavioural signals behind every threat score.
Traffic classification
The five traffic categories AdProtektor sorts every visit into — Real, Crawler, Bot, Competitor and Click Farm — and the finer states shown in the dashboard.
Tools and reports
The refund report, protection value report, keyword and search-term opportunities, ad copy help and AI traffic analysis.