AI agents for the insulation back office — that still sound like you.

AI agents for insulation contractors are assistants that handle the repetitive back-office follow-up an insulation business runs between jobs — checking in on unscheduled estimates, booking blower door tests after an install, tracking ConEd and NYSERDA rebate status, reminding customers about open balances, booking new-lead estimates with the right estimator, and requesting reviews after a job is paid and closed out. In FieldCamp they work because your estimates, jobs, crews, and rebates are real records: each agent reads the same data and only acts on what it can see, with tone control and an optional human-approval step so the messages still sound like your company.

The estimate follow-up agent watches estimates that have been sent but not yet scheduled. It sends a warm, plain-language check-in at +3 days, again at +7 days, and a final last-chance nudge — written in your tone and referencing the actual estimate, not a generic template. You can keep an approval step on it (it drafts, you tap to send) or let it run on its own. It knows when to stop following up so it never pesters a customer.

Yes. Because the job, its areas, and the rebate are real objects in FieldCamp, the BDT scheduling agent can see when a job moves to closed out and reach out to book the blower door test that gates the rebate payout. The rebate status agent tracks where each ConEd and NYSERDA filing stands and flags the jobs waiting on a single step — so a rebate never sits stalled because nobody booked the test or chased the paperwork.

No. Every agent writes in a tone you set — warm, brief, or formal — and references the real estimate, job, or balance it's following up on rather than a merge field. You can also keep an optional approval step on any agent, so it drafts the message and you review or edit it before it sends. The agents handle the repetitive steps; the voice stays yours.

Yes. Approval is optional and set per agent: turn it on for the ones you want to review first and let the trusted ones run on their own. Every message an agent sends is logged on the record it came from, you can turn any agent on or off at any time, and an agent never acts on data it can't see. It's your operation — the agents just handle the follow-up you'd otherwise do by hand.

Automation handles deterministic workflows — appointment reminders, on-my-way texts, status-change handoffs, scheduled reports. AI agents handle the conversational, judgment work that sits on top: writing a check-in that sounds human, deciding a rebate job is one step from payout and chasing the blower door test, knowing when to stop following up. Most insulation teams run both — see the insulation automation page for the workflow side.