/tags/agentic-ai/Recent content in Agentic AI on Yiwei ShiHugoen-usSat, 09 May 2026 01:20:40 -0400/posts/20260509-multi-turn-llm-workflow/Sat, 09 May 2026 01:20:40 -0400/posts/20260509-multi-turn-llm-workflow/<p>If you&rsquo;ve used the OpenAI, Anthropic, or Bedrock APIs to build something more sophisticated than a chatbot, you&rsquo;ve probably written an agent loop — code that lets the model call tools, receive results, and decide what to do next. I recently built one for a document analysis pipeline at work, and a few things surprised me. This post is a distillation of those lessons, using a generic example.</p> <h2 id="the-setup-a-four-tool-pipeline">The Setup: A Four-Tool Pipeline</h2> <p>Imagine you&rsquo;re processing a document. For each item the model identifies, you want to:</p>