ora
linkedin.com
success
Claude Code · Haiku 4.51:04
intent

What does linkedin.com do and who is it for? Explain it back to me.

12steps
64.8sduration
$0.8394cost
175,735tokens
13 steps1 reasoning step4 searches
home
docs
docs
search
/marketing-solutions
/jobs
search
business.home
business.home
search
/solutions
search
100%
on-site discovery
63%
reliability
38%
link following
path origin
  • previous resource38%
  • prior knowledge63%
insight

The agent successfully assembled a comprehensive explanation of LinkedIn by combining prior knowledge with web search results, since LinkedIn's public website did not expose pricing, positioning, or competitive differentiation in directly fetchable pages. The agent bypassed broken/redirected on-site URLs (/pricing, /about returned 404s) and relied on external third-party sources and its own knowledge to answer all four dimensions of the task (what, who, pricing, differentiation). The site itself was minimally navigable for this discovery task—the agent had to work around missing landing pages and reconstruct the answer from web search results and business subdomain content.

  • Steps [5], [8], [10] were web searches that returned third-party blog posts comparing LinkedIn pricing, features, and business model—none of this information was discoverable on LinkedIn's own public pages. The agent cited 8 external sources, not LinkedIn's site, as the basis for pricing tiers and competitive positioning.
  • Steps [1] and [2] attempted direct URLs (/pricing, /about) that both returned 404s, forcing the agent to route through web search instead. Step [3] and [11] successfully fetched business.linkedin.com pages (homepage and marketing solutions), but those pages contained only high-level marketing copy without detailed feature lists, pricing, or alternatives analysis.
  • Step [9] (/jobs) confirmed the jobs feature exists but returned a React-heavy SPA with minimal extractable text. The agent inferred feature categories from page metadata and navigation hints rather than from rendered content.
  • LinkedIn's public site does not publish structured information about its own pricing, tiered feature comparison, or explicit competitive positioning—critical information for the 'discovery-evaluation' task. The agent had to assemble this entirely from prior knowledge (63%) and web search (0% counted here, but 5 searches were run).
  • The 'succeeded_with_heavy_bridge' outcome reflects that the agent completed the task but required significant external research and prior knowledge to do so; LinkedIn's own site did not provide the answer in fetchable form.

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 →