Turn the agent from a monitor into an operator.
The autonomous layer that watches POS, kitchen, inventory and loyalty together and acts on what it sees.
- POS, KDS and Inventory set up — NUA needs live data from these to act on anything
- A decision on autonomy level: suggest-only, or act-automatically for low-risk decisions
- 1
Review data connections
Under AI Agent → Data Sources, confirm POS, KDS, Inventory and Loyalty are all reporting — NUA's forecasts and actions only cover connected modules.
- 2
Choose an autonomy level
Start in Suggest mode (NUA recommends, a human approves) under AI Agent → Autonomy, and move individual automations to Autonomous once you trust them.
- 3
Enable the automation library
Turn on the automations you want — demand forecasting, reorder suggestions, dynamic rostering — one at a time under AI Agent → Automations.
- 4
Set guardrails
Cap spend-affecting actions (e.g. auto-reorder limits, discount ceilings) under AI Agent → Guardrails so autonomous actions stay inside limits you set.
- 5
Turn on Voice POS (optional)
If you want natural-language ordering at the counter, enable it under AI Agent → Voice — see the Voice POS guide for device setup.
- 6
Check the audit trail
Every autonomous decision is logged under AI Agent → Audit Trail. Review it daily for the first week to build confidence before widening autonomy.
- Move automations to autonomous mode one at a time, not all at once — it's much easier to trust (or debug) a single automation.
- The audit trail is the fastest way to understand why NUA made a call — read it before overriding a decision.
- Forecasting accuracy improves fast in the first 2–3 weeks as it learns your venue's actual patterns; don't judge it on day one.
- NUA's forecasts look off for a new venue.This is expected in week one — forecasting needs a short history to calibrate. Suggest mode is the safer default until then.
- An automation isn't triggering.Check it's both enabled under AI Agent → Automations and that its required data source is connected.