<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Pandora&apos;s Lab — Blog</title><description>Field notes on turning promising AI experiments into dependable software.</description><link>https://pandoraslab.dev/</link><language>en</language><item><title>How to build AI agents that work: start with the simplest thing</title><link>https://pandoraslab.dev/blog/como-construir-agentes-de-ia/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/como-construir-agentes-de-ia/</guid><description>Most AI agent projects get canceled over cost, unclear value or missing controls. When you actually need an agent, what to build first, and how to shape it so it&apos;s dependable.</description><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>Agents</category><category>Method</category></item><item><title>How to evaluate an AI agent before you trust it with your customers</title><link>https://pandoraslab.dev/blog/como-evaluar-agentes-de-ia/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/como-evaluar-agentes-de-ia/</guid><description>An agent that&apos;s right 90% of the time fails far more than it seems. How to build evaluations for AI agents from real tasks, what to grade, and why to read the transcripts.</description><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>Agents</category><category>Evals</category></item><item><title>Tools for AI agents: design them as if for someone new on the team</title><link>https://pandoraslab.dev/blog/herramientas-para-agentes-de-ia/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/herramientas-para-agentes-de-ia/</guid><description>Tools are how an AI agent acts on your systems, and the part that changes its results the most. How to choose them, name them, describe them and make them hard to misuse.</description><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>Agents</category><category>LLM pipelines</category></item><item><title>Context engineering: what your AI agent sees at each step decides what it does</title><link>https://pandoraslab.dev/blog/ingenieria-de-contexto-para-agentes/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/ingenieria-de-contexto-para-agentes/</guid><description>An AI agent doesn&apos;t remember your business. It only sees what&apos;s in its context at that moment. How to decide what goes in, what comes out, and how to keep focus on long tasks.</description><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>Agents</category><category>Method</category></item><item><title>AI adoption in Latin American companies: where things stand in 2026</title><link>https://pandoraslab.dev/blog/ia-en-empresas-latinoamerica-2026/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/ia-en-empresas-latinoamerica-2026/</guid><description>What the 2026 studies say about AI adoption in companies across Latin America and Colombia, why the numbers disagree, and what the companies getting real value do differently.</description><pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate><category>AI in Latin America</category><category>AI strategy</category></item><item><title>How to implement AI in your company without getting stuck in the pilot</title><link>https://pandoraslab.dev/blog/como-implementar-ia-en-tu-empresa/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/como-implementar-ia-en-tu-empresa/</guid><description>A six-step guide to implementing AI in your business and automating processes with it, from choosing the right process to running it in production with measurable results.</description><pubDate>Fri, 30 Oct 2026 00:00:00 GMT</pubDate><category>AI strategy</category><category>Method</category></item><item><title>AI for small businesses in Latin America: where to start without a big budget</title><link>https://pandoraslab.dev/blog/ia-para-pymes/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/ia-para-pymes/</guid><description>Only 6.8% of small companies in the region use AI in their operations. How a small business can start with one process, measure the result and know when to bring in an AI consultant.</description><pubDate>Thu, 29 Oct 2026 00:00:00 GMT</pubDate><category>AI in Latin America</category><category>Small business</category></item><item><title>AI agents for WhatsApp: what to require before they talk to your customers</title><link>https://pandoraslab.dev/blog/agentes-de-ia-para-whatsapp/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/agentes-de-ia-para-whatsapp/</guid><description>An AI agent on WhatsApp can answer, book and quote, and it can also make up a price. What a business AI agent needs before it goes live.</description><pubDate>Wed, 28 Oct 2026 00:00:00 GMT</pubDate><category>Agents</category><category>Guardrails</category></item><item><title>Inside the lab, where agents get tested</title><link>https://pandoraslab.dev/blog/inside-the-lab/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/inside-the-lab/</guid><description>A tour of the experiments every agent goes through before it ships, from the writing test to the night watch.</description><pubDate>Tue, 29 Sep 2026 00:00:00 GMT</pubDate><category>Method</category><category>Evals</category><category>Agents</category></item><item><title>Evals are tests that admit they&apos;re fuzzy</title><link>https://pandoraslab.dev/blog/evals-are-fuzzy-tests/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/evals-are-fuzzy-tests/</guid><description>Unit tests catch what breaks. Evals tell you whether a change made things better or worse, on examples that look like real life.</description><pubDate>Sat, 26 Sep 2026 00:00:00 GMT</pubDate><category>Evals</category><category>Method</category></item><item><title>Structured output is the cheapest reliability upgrade</title><link>https://pandoraslab.dev/blog/structured-output/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/structured-output/</guid><description>Free-form text is where AI pipelines quietly break. Give the model a shape to fill and validate every answer against it.</description><pubDate>Fri, 25 Sep 2026 00:00:00 GMT</pubDate><category>LLM pipelines</category><category>Python</category></item><item><title>Write the job down before you write the prompt</title><link>https://pandoraslab.dev/blog/define-the-job/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/define-the-job/</guid><description>Most unreliable AI features fail before the first model call. A one-page brief is the cheapest fix there is.</description><pubDate>Tue, 22 Sep 2026 00:00:00 GMT</pubDate><category>Method</category><category>LLM pipelines</category></item><item><title>Red-team your own agent before someone else does</title><link>https://pandoraslab.dev/blog/red-team-your-own-agent/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/red-team-your-own-agent/</guid><description>A short checklist for attacking your own AI feature, and what each attack should run into.</description><pubDate>Fri, 18 Sep 2026 00:00:00 GMT</pubDate><category>Guardrails</category><category>Security</category></item><item><title>Watch it live, or it will drift quietly</title><link>https://pandoraslab.dev/blog/watch-it-live/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/watch-it-live/</guid><description>Models, data and prompts all change after launch. Monitoring is how you notice before your users do.</description><pubDate>Tue, 15 Sep 2026 00:00:00 GMT</pubDate><category>Monitoring</category><category>Production</category></item><item><title>Working from Bogotá on US hours</title><link>https://pandoraslab.dev/blog/working-from-bogota/</link><guid isPermaLink="true">https://pandoraslab.dev/blog/working-from-bogota/</guid><description>What working with a lab in Colombia looks like day to day, from time zones to languages.</description><pubDate>Thu, 10 Sep 2026 00:00:00 GMT</pubDate><category>Working together</category></item></channel></rss>