As of mid-2026, a non-developer operator on a 1-50 person team can build a production-useful internal-process agent in Notion (via the 13 May 2026 developer platform: Workers, External Agents API with Claude/Codex/Decagon, Database Sync, ntn CLI) or in ChatGPT (via the matured GPT Builder with Instructions, Knowledge, Actions and four sharing tiers) without a separate automation tool. The default permission scope of the resulting agent is the workspace-wide access of the human who built it (Notion: workspace-admin if the builder is an admin; ChatGPT: the builder's API key scope on configured Actions), which is the operator-scale shadow-AI risk. The safe-deploy playbook (three guardrails: scope to one data source not whole workspace; read-only first for a week of parallel running; human approval on customer-facing actions; plus the one permission-scope rule that the agent inherits the builder's access not the user's) closes the gap using only the tools' native configuration.

Claim is scoped to the operator-level capability of a non-developer 1-50 person team to deploy an internal-process agent safely on Notion or ChatGPT in 2026. Does not assert no-code building is universally appropriate for client-facing or transaction-authority work; asserts the playbook is the discipline that converts default-broad-scope shadow-AI exposure into deliberate, configured deployment. 45-day review cadence calibrated to the no-code-platform feature velocity (Notion's developer platform was a step change in May 2026 and is on a fast roadmap; OpenAI's Builder surface has shifted three times across 2024-2025). Trigger conditions: (1) Notion or OpenAI ships per-agent scoping as the default (rather than inherit-builder-access as default) — would move toward Partial because operator-side configuration burden shifts to vendor-side default; (2) a published small-team or agency incident specifically traceable to a no-code agent inheriting workspace-wide access — would confirm operational exposure; (3) no-code agent building proves not production-useful at this scale (significant friction, hallucination rates that defeat the operational case, or pricing changes that reset the calculus) — would move toward Partial or Not holding; (4) a new no-code platform (Salesforce Agentforce small-team tier, Microsoft Copilot Studio non-tenant tier, Zapier-class platform with native agent orchestration) materially reframes the build-vs-buy choice for 1-50 person teams — would change the structural shape of the playbook. Siblings: OPS-072 (the Notion AI agents hub upgrade-check), OPS-075 (the operator shadow-AI piece), OPS-061 (the delegation framework and the 90-second test the agent should pass before client work).