Most of what people are calling “agents” today are basically deterministic workflows with one or two LLM calls glued together. That is not an agent. That is a at best API pipeline.
So I am genuinely curious: are there any real examples of agents handling large, messy, multi-step workflows at scale? Not demos, not toy projects, not VC decks.
AIinyourAI•1h ago