AX
also called Agent Experience
How it feels to be an agent using your product.
The agent-facing counterpart to DX: whether your systems expose clean tools, unambiguous responses, predictable failures and enough structure for an agent to act reliably.
The unsettled question of the moment. Products are increasingly used by agents as well as people, and an API that is merely human-readable is not necessarily agent-usable.
The design constraints are different from human-facing ones: consistent naming across tools, unambiguous return shapes, errors a model can act on, and enough stabilty that the same call behaves the same way twice. A human can infer through an inconsistent interface; an agent develops a wrong belief and repeats it.
Naming that says when to use it
Two tools named get_user and fetch_user with overlapping behaviour.
The agent cannot tell them apart and picks at random. Merging them, or naming them for the situation rather than the mechanism, removes a decision the model was never given enough information to make.
- blogWriting effective tools for AI agents — using AI agents ↗
Treats tools as contracts between deterministic software and a non-deterministic caller: clear boundaries, token-efficient results, evaluation against realistic tasks.
- blogIntroducing the Model Context Protocol ↗
The original MCP framing: an open standard that replaces per-integration glue with a client/server contract for tools and data.
- standardModel Context Protocol specification ↗
The normative document: hosts, clients, servers, JSON-RPC, resources, prompts and tools — plus the security section on consent, user control and care around tool execution.
DX · Developer Experience
How it feels to build with a tool or platform.
MCP · Model Context Protocol
An open protocol for plugging tools and data into any agent.
Tool
A function the model is allowed to call to affect the world.
Adoption Curve
The path from a few enthusiasts to real usage.