omni-ai-optimizer

Omni AI Optimizer

Optimize your Omni semantic model so Blobby (Omni's AI assistant) returns accurate, contextual answers.

Tip: Use omni-model-explorer to inspect current AI context before making changes.

Prerequisites

export OMNI_BASE_URL="https://yourorg.omniapp.co"
export OMNI_API_KEY="your-api-key"

Requires Modeler or Connection Admin permissions.

API Discovery

When unsure whether an endpoint or parameter exists, fetch the OpenAPI spec:

curl -L "$OMNI_BASE_URL/openapi.json" \
  -H "Authorization: Bearer $OMNI_API_KEY"

Use this to verify endpoints, available parameters, and request/response schemas before making calls.

How Blobby Works

Blobby generates queries by examining:

1. Topic structure — which views and fields are joined
2. Field labels and descriptions — how fields are named
3. synonyms — alternative names for fields
4. ai_context — explicit instructions you write
5. ai_fields — which fields are visible to AI
6. sample_queries — example questions with correct queries
7. Hidden fields — hidden: true fields are excluded

Impact order: ai_context > ai_fields > sample_queries > synonyms > field descriptions.

Writing ai_context

Add via the YAML API:

What Makes Good ai_context

Terminology mapping — map business language to field names:

Data nuances — explain what isn't obvious from field names:

Behavioral guidance — direct common patterns:

Persona prompting — set the analytical perspective:

Curating Fields with ai_fields

Reduce noise for large models:

Same operators as topic fields: wildcard (*), negation (-), tags (tag:).

Adding sample_queries

Teach Blobby by example. Build the correct query in a workbook, retrieve its structure, then add to the topic YAML:

Note: When exporting queries from Omni's workbook, you'll get JSON with table, join_paths_from_topic_name, and sorts using column_name/sort_descending. Map these to YAML as follows:

• table → base_view
• join_paths_from_topic_name → topic
• column_name → field, sort_descending → desc
• Workbook JSON includes filters, pivots, limit, column_limit which you can include in YAML (though filter syntax requires consulting the Model YAML API docs directly)

Focus on questions users actually ask — check Analytics > AI usage in Omni.

AI-Specific Topic Extensions

Create a curated topic variant for Blobby using extends:

Improving Field Descriptions

Good descriptions help both Blobby and human analysts.

Adding synonyms

Map alternative names, abbreviations, and domain-specific terminology so Blobby matches user queries to the correct field. Works on both dimensions and measures.

Synonyms vs ai_context: Use synonyms for field-level name mapping. Use ai_context for topic-level behavioral guidance, data nuances, and multi-field relationships.

Optimization Checklist

1. Inspect current state with omni-model-explorer
2. Check AI usage dashboard for real user questions
3. Write ai_context mapping business terms to fields
4. Add synonyms to key dimensions and measures
5. Curate ai_fields to remove noise
6. Add sample_queries for top 3-5 questions
7. Improve field description values
8. Consider extends for AI-specific topic variants
9. Test iteratively — ask Blobby and refine

Docs Reference

• Optimizing Models for AI · Synonyms · Topic Parameters · Model YAML API · Omni AI Overview

Related Skills

• omni-model-explorer — inspect existing AI context
• omni-model-builder — modify views and topics
• omni-query — test queries to verify Blobby's output