askshepherd.ai
success
Claude Code
intent
What does askshepherd.ai do and who is it for? Explain it back to me.
7steps
57.2sduration
$0.0829cost
29,470tokens
8 steps1 reasoning step2 searches
home
docs
home
home
search
search
/companies/shepherd-3
67%
answer from your site
40%
answer efficiency
75%
followed site links
answer sources
- from this site67%
- from search results17%
- from agent knowledge17%
url discovery
- given in the task20%
- followed a link60%
- found via web search20%
insight
Agent successfully gathered and synthesized core information about Shepherd's purpose, target market, and differentiation from the homepage and Y Combinator search results. However, pricing information is absent from the public site—the agent had to note this gap and explain that contact with founders is required, limiting the completeness of the evaluation task.
- ›Homepage (step [0], [2]) contained sufficient product positioning and differentiators—no additional site navigation was needed or available. The agent typed the URL from memory rather than discovering it via site navigation, and the /pricing endpoint returned 404, confirming pricing is not publicly listed.
- ›Search results (step [5]) were essential to answer 'who founded it' and 'Y Combinator backing'—facts present on the Y Combinator company listing but not fetched (404 on step [6]). The agent relied on search snippet content, not the YC page itself.
- ›The site is minimal and discovery-unfriendly: no pricing page, no competitor comparison, no case studies or public reviews mentioned. The agent had to acknowledge missing information rather than construct answers from fragments.
- ›No structured data, llms.txt, or machine-readable metadata was discovered. The agent synthesized the answer by parsing unstructured homepage text and filling gaps with model reasoning about market positioning.
- ›The 'light bridge' friction outcome reflects that the agent had to guess/infer some context (target market specificity, differentiators vs. competitors) from homepage messaging rather than reading explicit comparative positioning or use case documentation.
Help the next agent finish on askshepherd.ai
Front builds the agent version of your site and shows which agents visit, what they came for, and what your pages couldn't answer.
or talk to us about agent readiness →