
SERVICE 03 / 05 · PREDICT · EXPLAINABLE AI AND FORECASTING
THE PROBLEM, IN YOUR WORDS
“Wecan’ttrust,ordefend,whatthemodelsays.”
Explainable AI and forecasting
We build forecasts where every number shows its reasons and its range, with the documentation to defend them.
DELIVERABLES
What you get
Precision agriculture and food · Insurers and lenders · Public agencies
- 01Forecasts with reasons for every numberEach prediction lists what raised or lowered it, in the units you use.
- 02Uncertainty rangesA range with every estimate, never a single number on its own.
- 03Model documentation and audit trailWhat data, which version, and why, for every prediction.
- 04Governance readinessRecords you can show a regulator, an auditor or a board.
What drives this forecast
East Field · t/ha
Model baseline7.3
Crop type: soybean-3.31
Regional baseline, Ontario+0.20
Soil organic carbon-0.08
Soil pH-0.05
Soil clay content-0.05
Other factors+0.03
Forecast4.1
HOW WE APPROACH IT
- 01Agree what a good forecast meansThe decision it supports, the error you can accept, and the test seasons.
- 02Build and test on unseen dataModels are scored on seasons or cases they never saw in training.
- 03Explain and documentReasons, ranges and records attached to every number.

SOLUTIONS THAT USE THIS SERVICE
What it delivers
Questions
Does explaining a model make it less accurate?
No. The explanation is calculated from the same model; it does not change the prediction.
Can this be funded?
Some Canadian programs, such as NRC IRAP, may cover parts of an AI project; we can help you scope it.
Sound familiar? Tell us where you are today.
Next service: AI assistants

