Business Users Are Now the Fastest-Growing Users of AI Coding Agents

The key signal is not broader Codex usage; it is the change in buyer persona. When automation shifts from engineers to line-of-business users, spend migrates from a discretionary dev-tool budget into workflow and operating budgets, which is structurally larger but much more sensitive to governance, auditability, and permissioning. That favors platforms already sitting inside identity, collaboration, and business-process surfaces — MSFT and NOW most obviously, with CRM/TEAM as secondary beneficiaries — because they can package "safe automation" as an extension of existing seats rather than a new category sale.

The pressure point is on vendors whose moat is complexity friction. If non-technical users can create useful automations without a dev backlog, low-end RPA and point-solution workflow vendors face substitution risk as simple tasks get absorbed into broader agent platforms. That is a negative read-through for PATH over a 3-12 month horizon, and a more subtle margin risk for model-centric tooling: if agent platforms become model-agnostic, the economic rent shifts from the frontier model layer to the workflow/orchestration layer. In other words, usage can rise while pricing power falls.

The contrarian miss is that adoption velocity does not equal monetization velocity. A lot of this usage is likely sandboxed experimentation, not production workloads with retention. The real catalyst is enterprise security approval: if CISO/CFO reviews slow rollout, the same-day automation narrative stays a demo story, not a revenue story. Watch next-earnings commentary on agent attach rates, governed deployments, and expansion within existing enterprise accounts; if those metrics do not inflect, the market is likely overestimating near-term revenue capture.