AI audit · written findings · sample

Sample Co.

Fictional 22-person commercial HVAC service company · On-site session with the owner, office manager, and service manager · Labor valued at $32 per hour over 48 working weeks

Summary

Sample Co. spends about 23 staff hours a week on five repetitive tasks, roughly $35,328 a year in paid time. Three of them are good candidates for supervised automation. Together they could give back an estimated $17,203 a year. Billing disputes should stay with a person.

Recommended first step: a supervised pilot on opportunity 1, after a short AI use policy, because three staff are pasting customer details into free AI tools today.

Where the hours go

Task	Who does it	Hours / week	Cost / year
Sorting and answering new inquiries	Office manager, dispatcher	9	$13,824
Re-keying technician job notes into invoices	Office staff	6	$9,216
Following up on open quotes	Service manager	4	$6,144
Writing maintenance renewal letters	Office staff	2	$3,072
Handling billing disputes	Owner	2	$3,072
Total		23	$35,328

Ranked opportunities

Value is the share of today’s hours we expect automation could give back, at $32 per hour. It is an estimate, confirmed during a pilot.

1
Sort new inquiries and draft replies
Routine requests are sorted and a draft reply is prepared. The office manager approves every message before it is sent. Emergencies and complaints go straight to a person.
Estimated value
$8,294 / year
Hours given back
~5.4 / week
Effort
Medium
Risk
Low
2
Re-key job notes into invoices
Depends on whether the field service app can export job notes. Confirm the export before scoping.
Estimated value
$4,608 / year
Hours given back
~3.0 / week
Effort
Medium to high
Risk
Low
3
Remind customers about open quotes
One reminder per quote, drafted for the service manager to approve. Quotes that stay quiet after one reminder go to a person.
Estimated value
$4,301 / year
Hours given back
~2.8 / week
Effort
Low
Risk
Low

Leave these with a person

Billing disputes. They need judgment and a relationship. Automation would save little time and add risk.
Pricing and estimating. Errors are costly, and the service manager’s experience is the value.

Tools and risks we found

Microsoft 365 email and spreadsheets, accounting software, and a field service app. All three are already paid for.
Three staff use free AI chat tools, sometimes with customer names and addresses. There is no written rule about what is allowed.
The field service app may export job notes. That would make opportunity 2 much cheaper.

Next 30 days

Agree a one-page AI use rule for staff, including what customer information never goes into AI tools.
Check the field service app’s export options.
Decide whether to pilot opportunity 1 with Turnrow, with another provider, or on your own.

Want findings like these for your business?

Book a free fit call