https://kalpakkale.com/blog/ Field notes on putting AI to work · evaluations, agent workflows, shipping without a dev team. en https://kalpakkale.com/blog/the-21-case-test-before-a-rag-chatbot-ships/https://kalpakkale.com/blog/the-21-case-test-before-a-rag-chatbot-ships/Fri, 04 Sep 2026 00:00:00 GMTA document chatbot for a regulated financial client went from a proof of concept to client beta. What got it there was not the model. It was a weighted evaluation of 21 cases, nine rounds of structured QA, and a daily accuracy report the client ran themselves. https://kalpakkale.com/blog/giving-19-developers-ai-through-one-proxy/https://kalpakkale.com/blog/giving-19-developers-ai-through-one-proxy/Fri, 28 Aug 2026 00:00:00 GMTRolling AI coding tools out to a whole team is not a licensing decision. It is a gateway, a spending policy, an evaluation habit and a set of working rules. Here is how a sixty-person services firm did it, including the 38 tool evaluations that mostly ended in "no". https://kalpakkale.com/blog/a-brief-goes-in-software-comes-out/https://kalpakkale.com/blog/a-brief-goes-in-software-comes-out/Fri, 21 Aug 2026 00:00:00 GMTAn operations manager with an agentic coding tool can take a statement of work to a production system, but only with a fixed process. The one I use, with the payment-first design and security defaults that a travel platform forced me to learn.