Chat-RPT is a conversational natural language interface for SAP-RPT. Instead of manually constructing prediction requests, you can ask questions in any natural language and the system will automatically classify your intent, build the appropriate query, run predictions, and return results with explanations.
Simulate changes to your data and predict outcomes without modifying the original dataset. Supports both categorical changes (e.g., swap a supplier) and numeric adjustments (e.g., reduce cost by 20%).
Example queries:
Add a new hypothetical record to your dataset and instantly get a prediction for the target column. The model uses your existing data as context to predict the new row's outcome.
Example queries:
Validate the model's prediction quality on your specific dataset. The system masks known values, runs predictions, and reports accuracy metrics so you can assess reliability before acting on results.
Example queries:
Ask questions about your existing data without running predictions. Get aggregations, distributions, comparisons, and summaries in natural language.
Example queries:
Continue the conversation with context-aware follow-ups. Reference previous results, refine scenarios, or drill deeper without repeating context.
Example queries: