Pricing with AI: comps, not guesses
Ask an AI "what's this worth?" with nothing else and it will give you a number. The number will sound reasonable. It may be invented. Pricing is the area where AI needs the most help from real data.
Start from real prices
The best evidence is what identical items have actually sold for recently, then what's listed now. Filter for the same model, same condition, same completeness. Five good comparisons beat fifty loose ones.
What AI is good for here
- Summarising the comps you found: the low, the middle, the high, and why they differ
- Spotting what the expensive ones have that yours doesn't — box, accessories, better photos
- Turning a spread into a strategy: Buy It Now at the middle with offers, or an auction with a sensible start
What to watch for
- Made-up comparables. If the AI quotes prices you can't click through to, ignore them
- Mixed variants. The Pro, the Mk II and the standard model often sell at very different prices
- Asking prices that never sell. A listing sitting at £400 for months tells you £400 is too much
A sane pricing routine
- Find five to ten matching sold listings
- Note condition and what was included
- Let AI summarise the range and suggest a price with a one-line reason
- Decide yourself — you know your stock, your costs and how fast you need the cash
The Price check tool in the FRANK area shows live asking prices for any item with a buy / risky / pass verdict — a quick first look at the car boot. #pricing #sourcing
- Question
- Active discussion
- Answer emerging
- Community solution
- Knowledge article