The problem
A landlord opens a condition report full of codes and has no idea what to do next, so they ring you. At 5 reports a week and 15 minutes writing each cover letter, that's 75 minutes a week, about 5.4 hours a month, plus the follow-up calls.
What the AI does
You write up the inspection as normal and complete the certificate yourself. Then you paste your observations into an AI assistant and ask for a short note to the landlord: what each observation means in everyday terms and what happens next. You check it matches your report before sending. The classification and the technical judgement stay with you.
A worked example
A two-man electrical firm in Manchester does about 5 landlord inspections a week. After completing each report, one of them pastes the observations into ChatGPT, Claude or Gemini and asks for a landlord note that lists the items in the order of urgency he gives it.
Before
15 minutes writing each cover note, and a few queries a week
After
3 minutes per note to check, with fewer explanatory calls
What it's worth
About 4.3 hours a month saved (12 minutes on each of 5 reports a week), worth around £194 if your time is worth £45 an hour.
How to set it up
- Finish and sign off the EICR yourself exactly as you do now; the AI never produces or changes the certificate.
- Give your AI assistant a standing instruction: plain English, no jargon, short, and never change the observation codes.
- Paste in your observations only, without the landlord's name or address.
- Compare the draft with your report line by line and correct anything that shifts meaning.
- Add your name and contact details, then send the note alongside the report.
What goes wrong
- An assistant can get a technical point wrong with complete confidence. You are responsible for what the landlord is told, so read every line.
- It must never decide or soften an observation code. Keep the code and wording from your report in the note.
- Don't paste personal data you don't need. Use a reference number instead of the landlord's name.