You are here:
Semantic Model AI Optimization (Beta)
Semantic Model AI Optimization (Beta) helps data teams prepare their Semantic Models for use with AI agents such as Concierge: Analytics Q&A. It evaluates a semantic model's quality, identifies inefficiencies, and offers guided resolutions to improve model reliability for AI-powered analytics.
This feature helps data analysts achieve clarity on what makes a semantic model "AI-ready" and provides targeted guidance on where to start and what to prioritize. In addition to diagnosing problems, the tool offers suggested improvements.
Use of this service might consume Customer Data Cloud Credits or Flex Credits.
How It Works
Semantic Model AI Optimization runs a series of automated checks against your semantic model and calculates an overall model health, which is categorized as Low, Moderate, or High. The assessment covers several key areas:
- Missing Description: Checks whether descriptions are needed in custom objects and semantic assets.
- Missing Relationships: Flags objects with disconnected relationships, as this can impair an AI agent's ability to reason across the model.
- Similar Meanings (Beta): AI helps identify semtnaitc definitions that are so similar, it's hard for both the system and dashboard designers to tell them apart. This can lead to inconsistent or random results in your analytics.
The assessment presents results in a panel that visualizes the Model Health strength for you to explore the rules, issues, and suggested resolutions. The feature also provides a contextual call-to-action button for each rule, tailored to the resolution path. Acting on these suggestions improves your model’s readiness for AI-powered agents and analytics.
What’s an AI-Ready Semantic Model?
An AI-ready semantic model is a data model structured specifically for AI comprehension. While traditional semantic models were designed primarily for human understanding and dashboard creation, AI agents require models with a high degree of precision, clarity, and consistency.
AI readiness in a semantic model means that AI agents can reliably interpret business context from it. An AI-ready model has clear labels, complete descriptions, low ambiguity, and consistent metadata. Without these elements, AI agents struggle to provide accurate output, which can slow AI adoption and erode user trust. Data teams use the Semantic Model Optimization feature to prepare their models by guiding them to focus on the highest-impact improvements needed for AI-driven use cases.
To learn more about semantic models and best practices for creating them, we recommend completing the Design Your Semantic Model for Analysis Trailhead.
- Enable Semantic Model AI Optimization (Beta)
To begin using Semantic Model Optimization, an administrator must first enable it. - Using Semantic Model AI Optimization (Beta)
After a system administrator enables the feature, you can begin using it to evaluate and improve your semantic models. You interact with the feature primarily through the Optimize Model Panel, which visualizes the Model Health and AI-readiness strength of your models. - Resolve Overlapping Data Definitions with Similar Meanings (Beta)
Improve model clarity by proactively resolving similar definitions that confuse AI agents. The Similar Meanings (Beta) feature uses AI to detect synonyms, acronyms, and semantic overlaps to prevent ambiguity. By analyzing field labels, descriptions, and parent object semantics, the tool identifies when a model contains "two names for one thing" or "the same name for different things"—concepts that often lead to inaccurate agent reasoning. This deep scan allows you to distinguish between closely related fields, helping agents correctly interpret your data and provide more reliable answers.
