Why should you care?

AIPredict dashboards and emails need different fraud tips vs upsell tiles per customer tier.

See it live — copy this example

Use a .NET console or Web API project with Microsoft.ML. Run dotnet run after pasting.

var segmentModel = ml.Transforms.Concatenate("Features", "Recency", "Frequency", "Monetary")
    .Append(ml.Clustering.Trainers.KMeans(numberOfClusters: 5)).Fit(rfmData);
var segment = ml.Model.CreatePredictionEngine<RfmRow, ClusterPred>(segmentModel)
    .Predict(currentRfm).PredictedClusterId;
var recModel = segment switch { 0 => highValueRecModel, _ => growthRecModel };

What happened?

• Cluster or score users, pick policy/model per segment, then serve personalized recs and messaging.
• Follow the steps below — typing the code yourself is the fastest way to learn.

Practice next

1. Build RFM features.
2. KMeans segments.
3. Route segment to rec model.
4. Add churn probability as a segment input.
5. A/B segment-specific layouts.

Remember

Segment → policy. Different models per tier. Personalized slots.

AIPredict tiered offers

VIPs see premium bundles; new users see intro deals.

Outcome: Conversion up without discounting everyone.