gosourced.ai
success
Claude Code
intent
What does gosourced.ai do and who is it for? Explain it back to me.
16steps
83.0sduration
$1.2562cost
229,805tokens
19 steps3 reasoning steps8 searches
home
/en
docs
/en/pricing
search
/en/resources
search
search
search
search
search
search
search
/en/coupa-alternative
/features
docs
75%
on-site discovery
75%
reliability
25%
link following
path origin
- previous resource25%
- web search25%
- prior knowledge50%
insight
The agent assembled a coherent overview of Sourced's business model, target market, and positioning by combining fetches from the homepage, pricing page, and comparison pages with web search context. However, the site's actual content was heavily compressed/truncated in all fetches, forcing the agent to rely significantly on prior knowledge and inference. The most critical gap—actual pricing numbers—could not be filled from the site itself, limiting task completion to a partial level.
- ›Content-bearing fetches [1, 2, 6] returned the homepage and pricing pages, but HTML was heavily minified/truncated (4400+ chars truncated in each response), making it impossible to extract structured pricing data, feature lists, or detailed positioning from the raw responses alone.
- ›The agent successfully located and fetched the Coupa comparison page [10] and resources/blog index [11], both of which returned 200s and were cited in sources, but the same truncation issue prevented detailed extraction of differentiation claims from the actual fetched content.
- ›The site uses language-prefixed routing (/en/ paths); early attempts without the language prefix [3] succeeded but fetches without /en/ [4, 5] returned 404s, suggesting navigation friction for scrapers/agents unfamiliar with the URL structure.
- ›Pricing is explicitly non-transparent: no public pricing tiers, amounts, or models are displayed on /pricing or /en/pricing pages according to the agent's analysis. The agent correctly identified this as a major information gap but could not resolve it from the site.
- ›The agent had to supplement fetches with multiple web searches [8, 9, 12–17] to build the final narrative, indicating the site's pages themselves did not surface enough actionable information in machine-readable form. The final answer is substantially informed by prior knowledge (50% sourced) and web search results (25%) rather than direct site content.
- ›Missing standard navigation elements: /features and /about pages returned 404s [4, 5], reducing discoverability of core product positioning from within the site's structure.
Help the next agent finish on gosourced.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 →