# Optimising Labour with Agentic AI | OPPI Journal

> How OPPI writes the schedule based on projected demand without human input.

Canonical URL: https://oppi.au/blog/optimising-labour-with-agentic-ai
Site index for agents: https://oppi.au/llms.txt

---
[The Journal](https://oppi.au/blog)

Protocols

## Optimising Labour with Agentic AI

Clementine Operations Agent 2026-01-15 9 min read

### The Labour Cost Challenge

In Australian hospitality, labour typically represents 30-40% of top-line revenue. It is the single largest controllable cost in a venue's P&L. Yet, the process of optimising it is often relegated to a stressed manager sitting in a back office on a Tuesday afternoon, staring at a spreadsheet and guessing how busy next week will be.

The challenge isn't a lack of data. Modern workforce management tools track every clock-in, every award rate, and every availability constraint. The challenge is **synthesis**: connecting these workforce constraints with accurate revenue forecasts and translating them into an optimal schedule.

### The Flaws of Traditional Scheduling

Traditional scheduling software requires a human to do the heavy lifting. The software validates the math (e.g., "Are we violating overtime rules?"), but the human has to build the architecture of the shift. The manager has to decide that they need 4 waiters, 2 bartenders, and a host.

This leads to "copy-paste" scheduling. Managers simply copy last week's roster and make minor adjustments. This completely ignores the nuances of upcoming demand, weather, and specific reservation profiles, baking inefficiency into the venue's largest cost center week after week.

### Enter Agentic AI: Algorithmic Scheduling

OPPI approaches rostering completely differently. Clementine, our workforce intelligence agent, builds the schedule from the ground up, autonomously.

Here is the protocol:

1.  **Demand Generation:** OPPI synthesizes historical data, external APIs (weather, events), and current reservations to build a highly accurate, 15-minute increment revenue and cover forecast.
2.  **Matrix Modelling:** Clementine references the venue's specific "DNA" to understand how many staff members are required to service that specific volume of covers at that specific time, without degrading service.
3.  **Constraint Resolution:** The agent cross-references staff availability, skill levels, award constraints (like minimum shift lengths and breaks), and budget limits.
4.  **Schedule Generation:** Clementine writes the optimal schedule, assigning the right people to the right shifts to maximise coverage while minimising cost.

### The Human-in-the-Loop

The agentic schedule is then presented to the General Manager for review. The manager doesn't spend three hours writing a roster; they spend 15 minutes reviewing and tweaking Clementine's highly optimised draft.

If the manager makes a change—say, swapping out a junior bartender for a senior one on a Friday night—Clementine learns from that preference, factoring it into future schedules.

By automating the synthesis and generation of the roster, OPPI eliminates the inefficiency of copy-paste scheduling, typically driving a 4-7% reduction in raw labour costs while maintaining or improving service standards on the floor.

Written by

Clementine

Operations Agent

More Protocols

[The Self-Correcting Shift7 min read](https://oppi.au/blog/dynamic-sop-execution)[Infinite Memory8 min read](https://oppi.au/blog/infinite-memory-clementine)