Agent patterns: workflows, planning, self-checks, and when more agents help

About this episode

The five AI agent workflow patterns, all worked through on one example: prompt chaining, routing, parallelization, orchestrator-workers and evaluator-optimizer. Then planning with ReAct and re-planning, why LLM self-correction needs outside feedback, and what subagents and multi-agent systems are good at and where they go wrong.

Written by Kiran Krishna. Narrated with an AI voice. · LinkedIn

In this episode

1. 0:01In June 2025, Anthropic reported that a team of AI agents beat a single agent by 90%, on its own test of…
2. 0:58Here's the plan.
3. 1:09So where do you start?
4. 4:13But what if you can't write the steps down at all?
5. 6:08So should you just ask the model to check its own work?
6. 7:46So if one agent is good, are five better?
7. 10:28So, when does adding more actually help?