ora
growthbook.io
success
Claude Code
intent

What does growthbook.io do and who is it for? Explain it back to me.

13steps
47.8sduration
$0.1288cost
122,346tokens
18 steps5 reasoning steps3 searches
home
docs
/features
/solutions
docs
search
/comparison
/compare/growthbook-vs…
/insights/optimizely-a…
search
/customers
search
mcp-call
90%
on-site discovery
60%
reliability
50%
link following
path origin
  • previous resource50%
  • web search10%
  • prior knowledge40%
insight

The agent successfully assembled a comprehensive explanation of GrowthBook by combining direct site fetches with web search. The homepage, pricing page, comparison pages, and customer stories provided the core narrative; however, several expected pages (features, solutions, comparison index) returned 404s, forcing the agent to rely on search results and prior knowledge to fill gaps. The site is moderately navigable but has structural gaps in its information architecture.

  • ›Step [1] (homepage) and step [2] (pricing) were the primary content sources, establishing GrowthBook's core product positioning (warehouse-native experimentation platform) and seat-based pricing model; both were fetched directly via guessed URLs and returned complete pages.
  • ›Steps [11] and [12] provided competitive differentiation detail (vs Optimizely, Statsig, LaunchDarkly) but had to be discovered via web search [9] rather than surfaced from the main site—the agent could not find a unified comparison page at /comparison or /features.
  • ›Step [7] (docs homepage) and step [16] (customers page) returned 200s but the agent could not extract detailed content from the HTML snippets; customer insights came from metadata descriptions rather than page body content.
  • ›The agent encountered three 404s on plausible paths (/features, /solutions, /comparison), indicating incomplete site structure; these forced reliance on web search results (steps [9] and [14]) to answer 'how it's different from alternatives' and 'who it's for'—information that should be natively findable on the site.
  • ›The agent correctly cited all fetched sources in its final bibliography, but ~40% of the answer relied on prior knowledge or external search results rather than direct site content, per the metadata note 'prior_knowledge: 40%, web_search: 10%, previous_artifact: 50%'.
  • ›Pricing details were cited from SaaSWorthy (external) rather than fully extracted from growthbook.io/pricing, suggesting the pricing page HTML was incomplete or required JavaScript rendering to expose detailed tier breakdowns.

Want to run your own?

Join the waitlist for early access to point your own agents at any domain, with the intents you choose.

or talk to us about agent readiness →