Choosing where to invest in a 115-person HVAC business.

The business problem

The business wanted a practical view of where AI could help before committing to implementation. Repair quotes, coordination, purchasing and contract renewals involved different people, information and decisions. They needed to be assessed as operational work, with realistic constraints.

My contribution

I worked through the opportunities and helped make the trade-offs visible. The assessment looked at the workflows, the likely value of changing them and the effort involved. It gave leaders an order of work and a basis for their next investment decision.

Trace the work from the handover

Repair quoting was one concrete example: an engineer records the problem, the office gathers the information, a quote is prepared and the customer decides. Mapping the handovers showed where information could be missing and where work waited for the next action.

Compare four opportunities on the same basis

The review covered repair-quote conversion, coordinator capacity, purchase orders and contract renewals. Looking at value and effort together made it possible to prioritise rather than treating every AI idea as equally useful.

Separate a proposal from a result

The plan considered where AI could prepare information and where a person should retain judgement. Those were proposed changes for the business to evaluate. The delivered output was the opportunity assessment and phased plan, rather than an implemented system.

What was delivered

• An operational opportunity map
• Four prioritised workflows
• A comparison of potential value and effort
• A phased plan for leadership

Results and scope

Leadership received a comparison of value and effort across four workflows, with a phased order for taking them forward.

Facing a similar problem?

For an MD with several operational bottlenecks, the first decision is where a change is worth the effort. We can discuss the process, information and people involved before deciding whether an implementation project is the right next step.

Bring one operational bottleneck and the decision you need to make about it.