Home/Our work/Onsen Hot Pools Rostering Automation + AI Agent
Onsen Hot Pools Rostering Automation + AI Agent
Live

AI Agent

Onsen Hot Pools Rostering Automation + AI Agent

An AI rostering assistant that drafts Onsen Hot Pools' weekly staff rosters, saving at least 5 hours a week for each of its three teams.

Completed June 2025

Key features

Automatic rosteringStaff and leave managementChat assistantHuman review built in

Onsen Hot Pools is an iconic Queenstown destination, offering several kinds of experience for travellers who want to relax with a view of the Shotover River.

Behind the scenes, matching staff to all those services took hours every week.

What we built

An AI rostering assistant that knows Onsen's staff, services and leave, and drafts rosters for the team to check.

  • Rosters drafted automatically: It follows more than 30 rostering rules and produces weekly rosters that are 98% complete, ready for a quick review.
  • Ask it in plain language: The team chats with the assistant to handle everyday staff tasks, or to get a roster on demand as a spreadsheet they can download and check.
  • Everything somewhere familiar: Staff, services, leave and logs live in Airtable, where the team can see and change them easily.
  • A person always signs off: The assistant drafts future rosters on a schedule, and someone at Onsen reviews each one.
  • What changed

    Rostering went from hours of manual work to a review. The assistant saves Onsen at least 5 hours a week for each of its three teams.

    It works because it fits into how the team already worked, with a clear point where a person checks the result.

    Built with

    n8n for the assistant and workflows, OpenAI, Airtable, and a custom rostering algorithm (simulated annealing) running as its own Node.js service on Railway. The assistant reaches its tools through MCP servers.

    About us

    We turn your goals into AI and software that actually works

    A team of product engineers based in Queenstown, NZ. We work with you to understand the problem first, then build the right thing — not just the possible thing.

    Book a consultation