The agent assembled a largely complete picture of Unblocked's product, pricing, and positioning from the homepage, pricing page, and context-engine page, but could not verify competitive differentiation from the site itself—it had to search externally to answer the 'how it's different' question. The site's own claims about alternatives were thin, forcing reliance on vendor-authored blog content and third-party listings.

• ›The site does not publish a competitors or alternatives page; comparison content exists only in Unblocked's own blog posts, so vendor bias is unavoidable. The agent had to search externally to answer 'how it's different from alternatives.'
• ›Performance claims (48% fewer tokens, 83% faster) are weakly supported: the 48% figure comes from a single undescribed engineering task with no sample size or methodology, and the 83% figure does not appear on the outcome page the agent checked—a discrepancy between the homepage claim and the supporting evidence.
• ›Pricing conflict: the agent found $19/user/month listed for 'Code Review' on one page but $29/$35 for 'Platform' on the pricing page, and could not reconcile the two from site content alone.
• ›Key technical terms ('context layer,' 'knowledge graph,' 'MCP,' 'RAG,' 'semantic retrieval') are used throughout but never defined on the pages checked, leaving non-specialists to infer meaning or consult external docs.
• ›The site omits company background entirely (no About page, founders, funding, or team visibility), making it difficult for a human or agent to assess credibility or organizational depth.