Know the right trade pricebefore you make the offer.
agentMO values a used car the way your sharpest dealer would — anchoring to live market comps, applying your trade rules, then correcting with an ML model trained on what dealers actually paid. Every number is explained.
For independent dealers and trade buyers — stop leaving margin on the table, or overpaying on a trade.
No sign-up. Runs entirely in your browser session.
Recommended trade-in
2021 Toyota Hilux 2.8 GD-6 Raider 4x4
85 000 km · Automatic · Cloth · Potchefstroom
R451 000
R72 000 gross headroom · 13.8% margin
Rule engine & ML agree · 23 comparable listings
- 99,337
- live SA listings
- 524
- model groups
- 40
- makes
- Real
- dealer trades
Three signals, one defensible number
How an agentMO valuation is built
Market comp pool
Pulls comparable on-variant listings from a snapshot of ~99,000 AutoTrader cars, then fits a km-adjusted regression so the baseline reflects this car's mileage.
Transparent rule engine
Applies the adjustments a dealer makes by hand — mileage, FSH, accident, colour, transmission, area, recon, target margin — with a book-value sanity anchor.
ML residual correction
A gradient-boosted model learns where the rule engine runs high or low versus real trades, then blends the two — weighted by confidence.
Paste a WhatsApp listing
Drop a dealer-group message in and the fields auto-fill — make, model, km, book values, FSH, recon.
See the why
Full adjustment breakdown, the comp pool behind the number, and the market price distribution.
Confidence, not false precision
Every valuation carries a confidence score and shows a range when the models disagree.
Price your next trade in seconds.
Manual entry or a pasted WhatsApp listing — you get a defensible number with the full reasoning behind it.