This is an illustrative scenario, not a real client. It shows how we'd approach a situation like this. Figures describe what we'd aim to measure, not results we're claiming.
The situation
A regional plumbing and heating firm, fifteen staff, wins work on responsiveness. The owner spends most evenings turning site notes and photos into written quotes, then chasing them. Quotes go out slowly, some never get followed up, and good jobs slip to whoever replied first.
Where AI could help
Two honest opportunities stand out, both well inside what AI does reliably today:
- Drafting quotes from structured site notes and a price list, so a first draft is ready in minutes for the owner to check and send.
- Follow-up prompts that flag quotes going cold, with a suggested message.
This is document drafting and reminders, not magic. The skilled judgement, what to charge, what the job really needs, stays with the owner.
How we'd approach it
- Intake. A short call to see how quotes actually get produced today.
- Opportunity scoring. We'd rank the quote-drafting idea against the follow-up idea on return, cost and effort.
- Proof of concept. We'd build a working demo that turns a handful of the firm's real (anonymised) past jobs into draft quotes, so the owner can judge the quality before spending on a build.
- Report. Honest cost modelling, including monthly running costs, and where it would and wouldn't help.
- Handoff. A development-ready brief if they choose to go ahead.
What success would look like
We wouldn't promise a number up front. In the proof of concept we'd measure the things that matter to this business: how long a quote takes to draft, how many drafts are usable with light edits, and how many quotes get followed up. If those don't move, we'd say so.