The problem
Survey fees are often set once and left, while the real time per job varies: a Level 3 on a Victorian terrace 40 minutes away takes far longer than one on a new-build round the corner. A sole surveyor doing 25 surveys a month without knowing hours per job type can under-charge a whole category for years without seeing it.
What the AI does
Export a year of jobs from your diary or job log with survey level, property age band, post code district, fee, travel time, site time and report time, and delete client names and full addresses. Upload the file to ChatGPT and ask for fee per hour, travel included, by survey level and property age, with a chart. Its data analysis runs the sums in Python, so you can ask to see the method. You decide which fees to change. Full data analysis is on Go (about $9£7€8A$13C$13 a month) and above.
A worked example
A sole surveyor in Exeter does about 25 surveys a month and logs fee, travel, site and report hours for each. ChatGPT shows that Level 3 surveys on pre-1920 houses earn the least per hour, so the surveyor adds $100£75€89A$144C$141 to that fee band.
Before
Pre-1920 Level 3 surveys priced like any other Level 3, about 5 a month
After
$100£75€89A$144C$141 added to the fee for those 5 a month: $499£375€446A$720C$705 a month more
What it's worth
About $499£375€446A$720C$705 a month in extra profit, against about $9£7€8A$13C$13 for ChatGPT Go. Adding $100£75€89A$144C$141 to about 5 pre-1920 Level 3 surveys a month is $499£375€446A$720C$705 and because the work itself doesn't change, almost all of it is profit, as long as bookings hold.
How to set it up
- Export a year of jobs as a spreadsheet: survey level, property age band, post code district, fee, travel, site and report hours. Delete client names and full addresses.
- In ChatGPT Go (€8, about $9£7€8A$13C$13 a month) or above, upload the file and ask: 'Work out fee per hour, including travel, by survey level and property age band. Show a table, a chart and the formula you used.'
- Work out two categories by hand to check it used the right columns.
- Ask follow-ups such as 'Which post code districts take the most travel per job?' and 'What would a $100£75€89A$144C$141 rise on pre-1920 Level 3s add over a year?'
- Update your fee table and the fees in your terms of engagement, then run the check again in 6 months.
What goes wrong
- It's only as good as your time log. If report hours are guesses, so is the answer.
- It may pick the wrong column or treat blanks as zero. Ask it to list the columns and rows it used.
- A higher fee can lose instructions to cheaper firms. Check local fees before raising them.
- Free data analysis is limited. Go and Plus are personal plans: turn off 'Improve the model for everyone' in Data Controls if you don't want your data used for training.