AI for construction and trades: organised calls and quote follow-up

When a call arrives while the team is on site, an identified assistant can record the type of work, area and timing and prepare a visit request or summary. The business defines acceptance rules and a person handles pricing and final scheduling. The effect is measured from the firm's own baseline.

Most builders and contractors do not have an office manager; they have a phone in a dusty pocket and a list of quotes they meant to follow up. An AI agent is the office you never hired. Review invitations use the same criteria for every eligible customer; private feedback and issue handling run separately. It has context on each job and each customer, so it is not a generic bot, it is the admin half of the business running itself while you are on the tools.

Bigger firms run Procore or Buildertrend; smaller contractors run Jobber, a spreadsheet, or a notebook. Wherever your jobs and quotes live, the AI layer reads from there to know which quote is outstanding, which appointment is tomorrow and which job just finished. The platforms with APIs connect directly; the spreadsheet-and-notebook setups work through a simple shared sheet or a lightweight CRM we set up. You do not have to adopt heavy project software to get the AI; if anything, the AI makes the lightweight tools you already use good enough.

Quotes go cold for one reason: nobody followed up. But following up feels pushy when you are doing it by hand a week later. An automated sequence solves both problems. Two days after the quote, a friendly check-in. A week later, a 'still interested, happy to tweak the spec' message. Two weeks later, a final nudge. It is polite, consistent and never awkward, because it is just the system doing its job. Builders who turn this on typically recover a meaningful slice of quotes they had written off, which on construction tickets is real money, not loose change.

We build your automations with Claude Code from Anthropic, which is why this costs what a tradesperson can actually justify rather than enterprise prices. You do not touch any of it; it runs in the background. ChatGPT from OpenAI is handy for the odd admin job (writing a tidy quote description, a customer update), but it is generic and does not talk to your customers live. The voice AI and the follow-up agents are built for your business and connected to your jobs. The only thing you need to know is that it answers the phone and chases the work, so you do not have to.

The business case is tested from the company's own baseline: eligible enquiries received, accepted visit requests, quote follow-up completed, staff time and errors. A pilot records these measures without treating an enquiry as revenue or promising payback.