All work

Workplace technology · Enterprise SaaS

Enterprise Workforce Reporting Copilot

An embedded reporting copilot that lets facility managers at a large workforce-management SaaS platform ask natural-language questions of workforce and booking data, no SQL required.

Live in production
01

The problem

The client, one of Australia's larger workforce management SaaS platforms serving top-tier Australian businesses, had workforce and booking data locked in a warehouse only technical staff could query directly. Facility managers and workplace strategists had to wait on a manual report or a SQL request for even a simple utilisation question, a pattern that didn't scale across a growing multi-tenant platform.

02

The solution

We took a working proof of concept to production: an agent built on the Model Context Protocol that composes calls across dozens of purpose-built analytics tools spanning bookings, resource utilisation, buildings and floors, and departments, backed by session memory for follow-up questions, SQL guardrails, prompt-injection protection, verified per-tenant isolation, and a feedback loop, embedded directly into the client's own web application as a native tab.

03

The outcome

The assistant is live in production across the platform's tenants, with cross-tenant isolation verified before go-live. The engagement has since expanded into a second phase adding inline charts and trends inside the chat window and a Super Admin analytics dashboard, with a further initiative already scoped.

What went into it

The build.

Model Context Protocol (MCP) agentic loopAzure OpenAIPython / FlaskReact / TypeScriptMulti-tenant isolation & SQL guardrailsFeedback-driven improvement loop

Work with us

Ready to put agents to work?

Book a short discovery call. We will tell you honestly where agentic AI fits your business, and where it does not.

Or talk to our AI agent, bottom right — built with the same technology we deploy for clients.