Enrichment Tool

Data enhancement allows you to apply AI tools to enrich your Knowledge Base with new insights, derived fields, and structured metadata. Use it to analyze unstructured content, extract themes, classify records, or append missing data.

Enhancements are applied at the Knowledge Table level and can be configured, previewed, and batched for safe deployment.

1. When to Use Enhancement Tools

Use enhancement when:

• Your existing content lacks structure (e.g. long meeting notes, raw text, scraped emails)

• You want to append new fields derived from AI analysis (e.g. sentiment, category, summary)

• You need additional metadata for filtering, ranking, or downstream processing

Common scenarios include:

• Classifying inbound queries by intent or priority

• Extracting action items from internal docs

• Enriching contact or company records from external data

2. How to Enrich a Knowledge Table

Step 1: Select the Target Table

Go to the Collections page and open the Knowledge Table you want to enhance.

Step 2: Open the Enrichment Panel

Click “Enrich with Tool” in the top-right menu. This opens the enhancement configuration interface.

Step 3: Choose a Tool

Pick a prebuilt enhancement tool from the library (e.g. Sentiment Classifier, Topic Tagger, Contact Lookup, or your own custom tool).

Step 4: Map Output Columns

Define how the tool's outputs should be saved into your table:

• Choose the output field (e.g. AI Sentiment)

• Map to a new or existing column

• Optionally overwrite existing values

Step 5: Run a Sample

Start with 5–10 rows. Review the outputs carefully. Look for:

• Format mismatches

• Unexpected content

• Missing or null values

Step 6: Run in Batches

Once validated, run the tool across the full table or in segments.

3. Use Case Library

Content Analysis

• Sentiment analysis of user or customer feedback

• Key phrase or named entity extraction from internal notes

• Topic classification of long documents

• Content quality or readability scoring

Data Enrichment

• Lookup of company and employee information

• Domain-to-location augmentation

• Contact or role enrichment from external providers

• Industry or sector classification

4. Best Practices

Safe Testing

• Always run enrichment on a small sample first

• Double-check output types and values before bulk execution

• Make sure the AI model or tool is appropriate for your field type

Performance Management

• Avoid enriching thousands of rows at once without validation

• Break large tables into batches

• Monitor processing time and resource usage

• Sync your knowledge collection after enrichment to re-index vector data if applicable

5. Notes on Sync & Post-Enrichment Behavior

After enrichment:

• Manual Sync May Be Required: Enriched fields are not automatically vectorized unless synced

• Re-indexing may change search behavior depending on what fields were added

• AI Optimized Fields should be reviewed if newly added fields are relevant for search