<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Capim Engineering</title><description>Behind the scenes of the engineering and AI serving thousands of dental clinics: technical decisions, mistakes, wins and real numbers.</description><link>https://tech.capim.com.br/en/</link><language>en</language><item><title>AI agent observability vs. traditional software</title><link>https://tech.capim.com.br/en/ai-agents/ai-agent-observability/</link><guid isPermaLink="true">https://tech.capim.com.br/en/ai-agents/ai-agent-observability/</guid><description>Why monitoring AI agents is a different game: traces, a single LLM evaluator, human review and prompts treated as code — our day-to-day routine.</description><pubDate>Mon, 27 Apr 2026 00:00:00 GMT</pubDate></item><item><title>Why humanizing an AI is a strategic decision</title><link>https://tech.capim.com.br/en/product-design/humanizing-ai-camila-identity/</link><guid isPermaLink="true">https://tech.capim.com.br/en/product-design/humanizing-ai-camila-identity/</guid><description>Designing the identity of Camila, Capim&apos;s AI assistant: from user research to AI-generated videos — and the 113 new clinics that connected.</description><pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Error analysis: from 40% to 95% accuracy in 2 weeks</title><link>https://tech.capim.com.br/en/ai-agents/error-analysis-ai-agents/</link><guid isPermaLink="true">https://tech.capim.com.br/en/ai-agents/error-analysis-ai-agents/</guid><description>How we doubled our agent&apos;s escalation accuracy with error analysis and manual annotation — and why LLM-as-a-Judge wasn&apos;t the way.</description><pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate></item></channel></rss>