Vienna-based OtterlyAI has launched Agent Analytics, a feature that reads a site's own server logs to show which AI agents visited, which pages they fetched, and how that crawling lines up with the brand's appearance in AI-generated answers. It was announced on August 13 via GlobeNewswire and had already been rolled out quietly to first customers, with 50+ websites tracking agent behaviour before the announcement.
The technical premise is the part worth noting: AI agents do not execute JavaScript, so a tag-based analytics tool cannot see them at all. GPTBot, ClaudeBot and PerplexityBot request pages every day and what they collect feeds the answers ChatGPT, Perplexity, Gemini and Google AI Overviews return — traffic that is entirely absent from a standard analytics dashboard. Server logs are the only place that activity shows up.
Agent Analytics sorts each visitor into one of three kinds, which is what makes the crawl counts readable rather than just large:
The pitch is a funnel — "from crawled to cited" — joining two signals that have been separate: whether an AI system read your page, and whether it then cited you. Citation reports now show crawl data for a brand's own URLs next to citation data, which lets a team tell a technical access problem apart from a content one.
OtterlyAI's own logs make the case concretely: Claude's crawlers spend most of their time on the pricing page while ChatGPT concentrates on the homepage. That is a difference no JavaScript-based tool could surface.
This is a vendor announcement, not independent measurement, and the framing is a marketing one — CEO and co-founder Thomas Peham calls "what's the ROI of AI search" the best question a marketing team can ask right now. Treat the funnel claim as a product direction rather than a demonstrated result: crawl-to-citation conversion rates and revenue attribution are on the roadmap, not shipped.
Setup runs through a Cloudflare Worker, a WordPress plugin, a Netlify Edge Function or a webhook, and historical log files can be uploaded for retroactive analysis. The company says it processes no personally identifiable data and is GDPR compliant. Content intelligence features are planned for later this quarter.
OtterlyAI was founded in 2024 by Thomas Peham, Klaus-M. Schremser and Josef Trauner, is fully remote, says it is used by more than 40,000 marketing professionals, and was named a Cool Vendor in Gartner's 2025 Cool Vendors in AI in Marketing report.
Sources: GlobeNewswire release, Setup docs, OtterlyAI
