The prompt is no longer the interesting part of an agent. The harness — orchestration, tool integrations, persistence, the rules governing agent behavior — is where all the complexity lives now.
Teams building agentic systems describe the same evolution. The first version was a structured LLM workflow: deterministic steps, clear inputs, clear outputs. It worked, but it was too rigid for what customers needed. The rewrite is fully agentic. The agent chooses tools, decides on follow-up actions, adapts to context. That autonomy requires a much more sophisticated harness to keep it from going off the rails.
Same dynamic in coding agents. The prompt is cheap. The harness — tool call management, persistence across sessions, multiplayer coordination, policy enforcement, observability — is the product. And the harness is where the context engineering burden concentrates.
I've argued elsewhere that controllability is not optional for enterprise teams. That requirement is enforced inside the harness, not the prompt. The prompt cannot be inspected, versioned, audited, or rolled back across sessions in any meaningful way. The harness can. So the harness is where governance lives, where memory lives, where the integrations with your real systems live.
If you are evaluating agent products, look at the harness. Look at the developer surface for adding tools, the controls for restricting them, the observability across runs, the persistence layer between sessions. The model is interchangeable. The harness is what you actually buy.
Sources
Related Essays
Controllability Is Not Optional. Enterprise Teams Do Not Want Magic
Enterprise teams do not want magic agents. They want control over which submodules load, which tools run, and what the agent remembers — because they have been burned by black boxes before.
Agents Are Software, and Software Needs a Factory
People talk about agent harnesses as if the harness is the interesting part. It is not. The interesting part is the factory — sandboxing, orchestration, persistence, model translation.
The Framework Trap
LangChain pivoted to LangSmith. E2B sells the sandbox. The agent harness is not the product — it is the thing you give away. Monetization lives in the infrastructure underneath.
Key takeaways
- Structured LLM workflows broke under real customer needs. Teams are rewriting them as fully agentic systems with adaptive tool use.
- Autonomy requires a much more sophisticated harness — orchestration, multiplayer coordination, policy enforcement, observability.
- The prompt is not where the differentiation lives. The harness is the product, and the harness is where the context engineering burden concentrates.
FAQ
Why are teams rewriting structured workflows as agentic systems?
Because deterministic step graphs were too rigid for what customers actually wanted. The rewrite gives the agent room to choose tools, decide on follow-up actions, and adapt to context — at the cost of needing a much more rigorous harness around it.
What does a sophisticated harness include?
Tool call management, persistence across sessions, multiplayer coordination between agents and humans, policy enforcement, and end-to-end observability. None of that lives in the prompt. All of it is engineering work.