Overview

The ML/AI Configuration page is where you train and manage the machine-learning models behind personalised predictions in Revenue Intelligence. It shows whether your organisation has enough data, the state of each model, and lets you train or retrain them as new data arrives.

Getting there

Open the ML/AI Configuration card on the Settings hub. The page now lives at /revenue-intelligence/settings/intelligence-coverage; the older /revenue-intelligence/settings/ml-ai address redirects there.

ML Readiness Status

The readiness panel reports an overall status — Optimal Data, Ready for ML or Building Data — and three signals:

Closed Deals — won or lost deals the models learn outcomes from.
Deals with Stage History — deals with stage progression, used for timing.
Months of Data — how much revenue history is available.

Each signal is labelled, for example Optimal, Ready for Win Probability, Ready for Forecasting or Building data. While data is still building, AI Coaching and Intelligent Nudges are available straight away.

Machine Learning Models

Each model shows its status (Active, Training or Not Trained), the data it has against its thresholds (for example 83/30), and when it was last trained.

Win Probability — predicts the likelihood of closing opportunities. Needs closed deals (won or lost): Required 15, Optimal 30.
Revenue Forecast — forecasts revenue trends using time-series analysis. Needs months of revenue history: Required 1, Optimal 12. A Model complexity line shows what your history supports, from “Simple average (limited accuracy)” up to “Optimal ML (best accuracy)”.
Deal Velocity — estimates time to close and identifies bottlenecks. Needs closed deals with stage history: Required 20, Optimal 40.

Training and retraining

Check the model has reached its Required threshold.
Select Train Model (or Retrain for a model that’s already trained). The button shows Training... while it runs, and a Training Queue panel lists what’s training and queued.
Select Refresh to update readiness after new deals close.

If a model is below its Required threshold, the page shows how much more data it needs (for example “Need 5 more closed deals”) and offers Force train anyway. Expect weaker predictions from a model trained this way.

Troubleshooting

“Need … more …”: the model is below its Required threshold. Wait for more data, or use Force train anyway knowing accuracy will be limited.
Predictions feel weak: aim for the Optimal threshold — accuracy improves as you move from Required towards Optimal.
“Error loading ML configuration”: select Refresh or reload the page.
Figures look stale: select Refresh.

ML/AI Configuration: Train Prediction Models