← Back to work
AI Agent2026

Airwaves

Airwaves are a UK-wide facilities management company that have grown rapidly over recent years. The internal process holding them together, receiving jobs by email, manually entering them into a third-party system, and chasing clients for missing information, had become the main constraint on further growth. We built the AI agent that removed it.

Airwaves

The challenge

Airwaves were handling a very manual process. New jobs would come in via phone call or email, and had to be manually reviewed, back and forth between the client for the additional details that were needed before being manually added to their job management system.

Information was frequently missed, meaning engineers who were to complete the job would arrive on site without all the information they needed, triggering further back and forth with client’s mid-job, often by the engineers themselves, costing time and money.

This manual process was a real ceiling to the growth that Airwaves wanted to achieve year on year and was one of the main obvious blockers for their growth. The administrative burden was becoming significant, error-prone and was only getting worse as the company continued to grow.

Our approach

The approach we chose wasn’t the easiest. Their job management system an off the shelf solution, so understanding exactly what we had access to, and how we could get access to it was the first step. Without understanding this part, we’d have to change approach entirely.

After understanding what was possible, the approach became much easier. We designed an AI agent that would read the incoming job emails, decide if the information supplied by the client was enough; based on the thousands of data points that Airwaves already had from operating over the last 10+ years. If it was enough, the job was categorised, tied to the correct location and be pushed into the job management system for scheduling. If it wasn’t enough, the AI agent would go back to the client automatically, requesting additional information before pushing the job into the system.

This all happens quicker than a member of the front office staff could even read the email!

The outcome

The potential of this agent was blisteringly obvious as soon as the agent and automation had been deployed within the organisation for user acceptance testing. We took a subset of jobs initially (roughly 1/100, scaling up to 1/10) to ensure the agent was working as expected and any issues could be ironed out.

This was successful, and the job intake via email part of front office is now fully automated. Administrative staff no longer spend their day on data entry and are freed up to spend their time on client work that genuinely needs a human tough.

Jobs are reaching engineers faster, and we’ve seen a 98% reduction of on-site information requests from engineers, meaning fewer delays, fewer frustrated clients and fewer wasted journeys!