Skip to main content

Overall Recommendation

Overall Recommendation is friction AI’s top-level metric for AI-driven discovery. It summarizes how strongly AI assistants recommend your brand across tracked prompts through three key drivers: Visibility, Sentiment, and Purchase Intent.

Definition

Overall Recommendation is a composite score that summarizes how strongly AI assistants recommend your brand in purchase-intent and consideration queries, relative to your competitors, based on tracked prompts over time.
In plain terms: Overall Recommendation shows how likely AI is to recommend you when customers ask what to buy.

What It’s Made Of

The Overall Recommendation score is built from three components that together capture both presence (are you in the answer?) and preference (are you the recommended choice?).

Visibility

How often your brand is mentioned in AI responses

Sentiment

How positively your brand is described when mentioned

Purchase Intent

How often your brand is suggested as an option to buy or choose

Why These Components

These three dimensions cover the full spectrum of what makes a recommendation compelling:
  • Visibility ensures you’re part of the conversation
  • Sentiment ensures you’re portrayed positively
  • Purchase Intent measures direct recommendation strength

How It’s Calculated

The Overall Recommendation score is computed as a weighted combination of the three components above. Weighting is optimized to reflect recommendation outcomes, not raw mentions.
The exact weighting prioritizes signals that correlate with actual recommendations over simple visibility metrics. A brand that appears rarely but is consistently recommended will score higher than one that’s mentioned often but never suggested as a choice.

Scope and Granularity

Your Overall Recommendation score can be viewed at multiple levels:
See your score across each AI platform:
  • ChatGPT (OpenAI)
  • Claude (Anthropic)
  • Gemini (Google)
  • Google AI Overviews
  • Perplexity
Your Overall Recommendation score is an aggregate across all tracked prompts. Your scores by category or platform may differ significantly. A strong overall score can mask weaknesses in specific areas.

How to Interpret Changes

When Your Score Goes Up

AI is mentioning your brand more frequently across prompts.
The language AI uses about your brand is more favorable.

When Your Score Goes Down

AI’s portrayal of your brand is less favorable than before.
You’re appearing less in queries where purchase decisions happen.
New prompts were added or removed, changing the sample.

Edge Cases and Limitations

AI responses aren’t deterministic. friction AI reduces noise by running prompts repeatedly and aggregating results over time.
Your Overall Recommendation score reflects what AI actually says today, not guaranteed future answers. Model updates can shift results.
Results can shift due to AI model updates, changes in browsing/citation behavior, or seasonal query intent.

Examples

Example A: High Visibility, Low Recommendation

Situation: Your Visibility score is 75, but your Purchase Intent score is 40. What it means: AI mentions you often, but rarely suggests you as the “best” option or recommends you directly. Fix focus:
  • Product comparisons and “best for” content
  • Review aggregation and social proof
  • Clear use-case positioning

Example B: Low Visibility, High Sentiment

Situation: Your Visibility score is 35, but your Sentiment score is 85. What it means: When AI does mention you, it’s very positive. But you’re absent from many prompts entirely. Fix focus:
  • Content coverage and entity optimization
  • Citation building and distribution
  • Expanding presence in new query categories

FAQ

No. Share of voice typically measures raw mention frequency. Overall Recommendation is weighted toward recommendation outcomes, meaning it prioritizes being suggested as a choice over simply being mentioned.
Each AI platform (ChatGPT, Claude, Gemini, etc.) uses different models, training data, and citation behaviors. It’s normal to perform better on some platforms than others.
Scores update based on your analysis schedule. System prompts run automatically, and scores reflect the most recent data from tracked prompts.
Yes. friction AI tracks the same prompts for your competitors, so you can see exactly how your Overall Recommendation score compares to theirs.

Next Steps