ixp
latest
false
- Overview
- Introduction
- Extracting data from unstructured documents
- Building and deploying models
- Model building
- Model validation
- Model deployment
- Frequently asked questions
Important :
Communications Mining is now part of UiPath IXP. Check the Introduction in the Overview Guide for more details.
PREVIEW
Unstructured and complex documents user guide
Last updated Jul 2, 2025
Building and deploying models
This section contains a summary of the project creation and deployment process within the Unstructured and complex documents capability. Each section is detailed in the rest of the guide. Check the following overview based on each stage:
- Upload sample documents.
- Define the document taxonomy - Identify the specific data points (fields) and their relationships (field groups) you wish to extract.
- Build the taxonomy - Configure the extraction schema and provide comprehensive instructions to inform extraction predictions.
- Review LLM predictions - Assess the initial predictions to gauge model performance.
- Modify prompts - Adjust prompt instructions based on performance reviews and test their impact.
- Validate extractions - Confirm or correct extractions to collect accurate ground truth data for further validation.
- Review performance - Review the performance statistics for each model version.
- Compare models - Evaluate different model versions, including field-level performance, to ensure the model performs well for the intended use case.
- Refine performance to meet business acceptance criteria as required.
- Publish models - Once the models achieve the desired performance level, publish each model version for operational use and add deployment tags. Then, deploy models to make them available in Studio workflows.
- Roll back to a previous model version, if necessary.
