PRACTICAL AI · OCTOBER 7, 2026
How plumbers can turn field notes into a customer-ready job summary
The owner question
Can AI turn a plumber’s rough field notes into a clear customer summary without adding details that were never documented? Yes, if the task stays narrow: organize approved notes, preserve uncertainty, and send nothing until a person who knows the job reviews it.
Use three buckets
Give the draft a simple structure: work completed, observations or open questions, and next steps. Keep “observed” separate from “diagnosed.” If the technician wrote “corrosion visible near shutoff,” the draft should not upgrade that note into “the shutoff valve is failing.” Missing facts belong in a review list, not in the customer message.
A practical example
Fictional field notes: replaced the owner-approved kitchen faucet cartridge; tested hot and cold water; no active drip observed after testing; corrosion visible near the shutoff; no additional work approved; office will call about an optional follow-up visit. A useful draft reports exactly that. It does not claim the repair will prevent future leaks, diagnose the shutoff, add a warranty, or say a follow-up is booked.
Copy-and-paste prompt
Turn the verified plumbing field notes below into a customer-ready job summary with three headings: Work completed, What we observed, and Next step. Use only the supplied facts. Preserve words such as observed, possible, and not confirmed. Do not add a diagnosis, code conclusion, safety assurance, price, warranty, work not performed, or appointment. If a fact is missing or unclear, put it in a separate Owner review list. Keep the customer summary under 140 words. Field notes: [PASTE SANITIZED, APPROVED NOTES]. Approved next step: [PASTE NEXT STEP].
The 60-second human review
Before sending, compare every sentence with the original notes. Confirm the customer and job, work actually completed, parts used, test results, unresolved items, price or warranty language, and the real next step. Remove anything the record does not support. Route technical, safety, code, gas, warranty, and pricing questions to the person authorized to answer them.
Why the review boundary matters
Generative AI can confidently produce false or unsupported content, which NIST calls confabulation. If a job summary later becomes marketing copy, federal truth-in-advertising guidance also requires claims to be truthful, non-deceptive, and supported. AI can organize the wording; the plumbing business remains responsible for the facts and the decision to send it.