Enterprise agent lesson: permissions don't guarantee correct business transactions
shashib · x · 2026-10-08
A recap of a well-attended Infor Velocity Week session by Natalia Ptaszek (Infor) and Deepak Kovvuri (AWS) on the enterprise agent stack:
- The model is only one layer: Amazon Bedrock supplies model access and guardrails, while Infor's Industry Cloud Platform carries the applications, data, and business context.
- Planning and execution are separate: a supervisor agent plans the work; role-based agents execute via tools using the signed-in user's own permissions — if you can't change a delivery date, neither can your assistant.
- Safety ≠ correctness: an agent can hold valid permissions and still get the business transaction wrong, so permission enforcement can't replace business-level validation.
More from coding & agent
- Dev amazed as AI sets up a fresh computer's config in seconds, not minutes — Dimillian · 2026-10-08
- After a $10,811 Cloudflare bill, dev ditches delayed billing alerts for a Grok-built 3-hour usage monitor — mazzaTalk · 2026-10-08
- Voice agents that call businesses for people: caller ID, disclosure, billing and other pitfalls — Born-Mastodon443 · 2026-10-08
- Harvard study of 718 firms finds AI coding boosts individual output ~30% but firm-level shipping doesn't move — DavidLinthicum · 2026-10-08
- Dimillian shares a 4-step Codex workflow for native-looking iOS apps — Dimillian · 2026-10-08
- WSJ columnist declares 'the year of Linux' is finally here, thanks to AI agents — alexvoica · 2026-10-08