Powered by Zoomin Software. For more details please contactZoomin

Progress Data Platform Release Notes

NO TITLE FOUND

  • Last Updated: November 19, 2026
  • 3 minute read
    • Documentation

Enrichments process the resources in a dataset so that they can be searched, explored, and used in answers. Each dataset has its own enrichment settings, which apply to the resources that are ingested into it.

Enrichment options can be configured when creating a dataset or modifiyed and reprocessed at any point afterwards. See Change enrichments and Reprocess existing resources for details.

Enrichment options

The following table describes each enrichment and the Progress Data Platform functionality that depends on it. Enable the enrichments that support how you plan to use the dataset's content. If an enrichment is disabled, the functionality that relies on it might return less complete results or be unavailable for that dataset.

Notes:

  • When you create a dataset, all enrichments start disabled. You can create a dataset without any enrichments and enable them later.
  • Enrichments other than Chunking depend on chunking. When you select one of them, Chunking is enabled automatically. If you clear Chunking, the dependent enrichments are also cleared.
  • Enhanced Table Extraction, Fact Extraction, and Classification have additional settings. Click Settings icon next to the enrichment to configure it.
Enrichment Description Related Functionality Uses LLMs?
Chunking Splits resource text into smaller sections. Chunks support more precise search and are required by the other enrichments. Provides better assistant answers and Assisted Writing evidence. Enables the Chunks tab for resources. No
Vector Embeddings Creates vector representations of text chunks, enabling semantic search when a user's wording differs from the wording in a resource. Enables semantic search for assistant answers and Assisted Writing evidence. Without embeddings, search matches keywords only. No
Knowledge Graph Detects entities and relationships in resource text and links them to the project's knowledge graph. Links resources to entities in the project's ontology. Enables the Knowledge tab for resources and business data in graph views and graph queries. Yes
Enhanced Table Extraction Extracts tables from resources so that table content can be searched and analyzed. Improves search results and answers that rely on table content. Yes
Fact Extraction Extracts values for configured facts and stores them on resources. Makes extracted facts available for agent queries and dashboards. Provides publication details, such as titles and authors, for Assisted Writing citations. Extracted facts can be used in SQL and SPARQL queries. Yes
Classification Matches resources and text chunks to concepts in project taxonomies and stores the labels as searchable metadata. Enables filtering search results by taxonomy labels, and adds classification context to assistant answers and Assisted Writing evidence. Classifications can be used in SQL and SPARQL queries. No

Change enrichments

To change enrichments for a dataset:

  1. Open the dataset and select the Enrichments tab.
  2. Enable or disable enrichments; then, configure them as needed.
  3. Click Save Changes. To undo changes that you have not saved, click Discard.

Saved settings apply to content that you ingest afterward. They do not change the enrichment output of resources that are already in the dataset. To apply enrichments to existing resources, see Reprocess existing resources.

Reprocess existing resources

To apply changed enrichment settings to resources that are already in the dataset, reprocess them:

  1. Open the dataset. If the settings have changed since the last reprocess, Data Platform displays a banner with Reprocess now.
  2. Click Reprocess now in the banner, or click Reprocess All in the dataset's Actions panel.
  3. In the confirmation window, review Enrichments to run. All active enrichments are selected by default. Clear any enrichment that you do not want to regenerate. At least one enrichment must be selected.
  4. Click Reprocess.

The run is queued and might take time to finish. You can review its progress in the Enrichment Runs panel on the Enrichments tab.

To reprocess a single resource, open it from the dataset's Explorer tab and click Reprocess.

Alert