57% Ship Multi-Stage Agents. Only 16% Cross a Team Boundary.
Anthropic's 2026 enterprise survey shows most agents now handle multi-step work inside one team. The wall isn't building the agent — it's the jump to cross-functional, where integration and data access decide everything.
NeuroX AI · July 11, 2026
Anthropic's new enterprise report (How enterprises are building AI agents in 2026) has one number that quietly redraws the map. More than half of organizations — 57% — now deploy agents for multi-stage workflows. Only 16% run cross-functional processes across multiple teams. That 41-point gap is the whole story.
Building an agent that handles a multi-step task inside one team is now table stakes. It reads a ticket, calls a few tools, writes a result — all within a boundary where the data is clean and one owner catches the failures. The hard jump is the next one: an agent that moves work between finance and ops, or support and engineering. That's where the survey's top blockers live — integration with existing systems (46%), data access and quality (42%), and change management (39%).
None of those are model problems. They're the same production discipline that decides whether a prototype survives contact with real systems: where the seams are, who owns the data on each side, what happens when the agent hands off and the receiving system says no.
The single-team agent is a demo of capability. The cross-functional one is the business case — and 84% of organizations haven't crossed it yet. The gap isn't a smarter model. It's the wiring.