Refunds, fraud signals and high-stakes calls: where should agent autonomy stop?
ConvertMyStore · reddit · 2026-09-22
The author argues that what an agent can do differs from what it should be authorized to do. Refunds are a good case: agents can identify orders, check policy, and compute amounts, but ambiguous delivery evidence, fraud indicators, and financial exposure still demand human judgment.
His proposed boundary: clear evidence + low impact → automate; ambiguous evidence → review; high impact or policy exceptions → human approval. He asks the community which actions they would never let an agent execute autonomously, even when technically capable.
More from coding & agent
- User Vibecodes Tampermonkey Scripts to Patch Long-Ignored UX Bugs in University Systems — craig_macdonald · 2026-09-22
- Speaking UI layouts into existence with Jev, shadcn and a local transcription model — _AustinCalvert_ · 2026-09-22
- ai-copywriter hits 1k stars: an agent skill that writes converting copy with zero AI tells — tom_doerr · 2026-09-22
- ContextBridge: an open-source pool that routes AI tasks across local machines, APIs and shared hosting — IamAngusU · 2026-09-22
- Microsoft's 44-page playbook: 100+ internal AI projects show licenses alone fail — alex_verem · 2026-09-22
- signal-mcp: open-source local-first MCP server exposing Signal to AI clients — googlarz · 2026-09-22