Model Customization

The Model Customization feature in Predict4Dynamics empowers CRM administrators to take full control over ML-driven predictions, ensuring that insights are perfectly aligned with your organization’s unique business logic and data.

This feature allows administrators to modify shipped prediction models by selecting the most relevant columns, applying filters, and adjusting parameters to reflect real-world business scenarios. This ensures that every prediction—whether for leads, opportunities, customer churn, or project outcomes—is tailored to the context and priorities that matter most to your team.

For example, with Model Customization, you can predict sales forecast for “leads that are most likely to convert this quarter based on region and engagement history”. The tool dynamically applies your custom configurations to generate accurate, actionable forecasts, helping sales, service, and operations teams make informed decisions faster.

Step to customize AI Prediction model:

• Open the Predict AI Model record.

• In the configuration window, define the data for training by adding or editing the FetchXML query under “Include records that meet the following condition.”

• From the “Select columns that influence the prediction” list, choose the fields that impact the prediction (e.g., Rating, Annual Revenue, Job Title, etc.).

• Finally, select the target column under “Select a column to predict” and save the configuration to apply changes for the next model training, and then click on the configure model button to start the model training with the updated data

By combining user-driven model adjustments with the power of AI, Predict4Dynamics enhances prediction accuracy, improves interpretability, and ensures that insights are aligned with your organization’s unique objectives. This feature not only saves time but also boosts confidence in AI-driven decision-making, transforming raw data into a strategic asset across Dynamics 365 CRM.