Create, Connect, and Activate Models

Create your own predictive models with clicks, not code. Establish connections to bring in generative and predictive AI from models outside of Salesforce. Then, activate models to use the inferences and insights revealed in your data.

• Access Model Runtime Settings
  The new default runtime is the foundation for building all new predictive models in the AI Models tab in Data 360. This enhancement replaces the legacy runtime as the long-term platform for expanded predictive modeling capabilities.
• Create Predictive AI Models From Scratch
  Create models to analyze your data and get AI powered predictions based on machine learning. Learn about supported use cases and ethical model building. Then, follow the steps to create a model and evaluate model quality.
• Structured Clustering (Beta)
  Use clustering in Data 360 to uncover hidden patterns and discover natural groupings in structured data based on similarities in the fields that you select. Clustering identifies emerging customer profiles, recurring support patterns, and unexpected behavioral trends.
• Forecasting
  Use forecasting in Data 360 to predict future values from historical data. Forecasting models automatically identify trends and seasonal patterns to generate predictions without requiring data science expertise.
• Sentiment Analysis
  Use sentiment analysis in Data 360 to evaluate unstructured text and learn how customers feel about your brand, product, or service. For each record, such as a customer review or survey response, the model assigns a positive, negative, or neutral label and a confidence score. To filter audiences, build dashboards, and trigger automations, apply the model outputs.
• Topic Classification
  Use topic classification in Data 360 to organize unstructured text such as reviews, case comments, or survey responses, into meaningful categories. To evaluate records for classification, the model applies a trained language model to a selected field. For each record, the model assigns a topic label and a confidence score. Data 360 adds these outputs to your dataset. Apply topic classification in reporting and queries.
• Predictive Model Monitoring
  Maintain the accuracy, stability, and reliability of your deployed predictive models with model monitoring in AI Models (formerly Einstein Studio). This ongoing process provides in-depth insights into model's performance and identifies data quality or connectivity issues. Without consistent monitoring in Data 360, model performance can decline and lead to inaccurate predictions, unexpected outcomes, and reduced confidence in your results.
• Connect a Model
  Connect predictive and generative AI models hosted outside Salesforce and use them with Salesforce data to predict future outcomes. Select your data source, define your predictive criteria, and train your model to reveal predictive inferences and insights using AI and machine learning.
• Activate Your Model
  After saving your predictive model settings, the model is ready for activation. To consume predictions, you must first activate the model.
• Predictive Models in Companion Orgs
  Access predictive models trained and activated in a home org from your Data Cloud One companion org. Data Cloud One supports multiorg access, which is ideal when a business is spread across different teams, countries, or brands. Model governance in the home-organization also simplifies model management and enables you to scale predictive AI across your entire ecosystem.
• Model Lineage in AI Models
  Model lineage provides a traceable history of an AI model's lifecycle, from training data, versioning, deployment, and its use with live data. Data 360 integrates this capability in AI Models (formerly Einstein Studio) and Governance.
• Manage Large Language Model (LLM) Access by Hiding Configurations
  Hide an LLM configuration when you want users in Prompt Builder to change to a different model in their prompt templates. They can select a newer LLM version or a different LLM, possibly from a different provider. You can suspend access to a new LLM until you've completed testing. Optionally, you can later show a hidden LLM configuration to make it visible for selection.