Why measuring AI impact is different
Traditional SEO metrics (rankings, clicks, impressions) are not sufficient for AI-search. AI tools give direct answers – you measure success based on mentions and citations, not rankings.
The key KPIs for AI-search
1. Mention Rate
Definition: The percentage of prompts from your test set where your brand is mentioned.
How to measure: Systematic prompt testing (see testing and optimization)
Target:
- Establish baseline in month 1
- 10-20% improvement per quarter is realistic
2. Citation Rate
Definition: The percentage of mentions where your website is cited as source (with link).
How to measure: Prompt testing in Perplexity (shows sources explicitly)
Target: 50%+ of mentions should also be citations
3. Share of Voice
Definition: How often you vs competitors are mentioned in your category.
How to measure: Test comparative prompts ("best [service] agencies")
Target: Top 3 position in your core market
4. Correctness Score
Definition: Percentage of correct information in AI answers about your brand.
How to measure: Manual verification during prompt testing
Target: 90%+ correctness
5. AI Referral Traffic
Definition: Traffic from AI tools to your website.
How to measure:
- Google Analytics 4: filter on referrers
- Look for: chat.openai.com, perplexity.ai, etc.
- UTM tracking where possible
Note: This is often limited measurable because many AI answers don't generate clicks.
6. Assisted Conversions
Definition: Conversions where AI-search played a role in the customer journey.
How to measure:
- Post-conversion surveys: "How did you find us?"
- Multi-touch attribution
- Mention monitoring + conversion correlation
Tracking setup
Manual tracking
Prompt testing spreadsheet
- Monthly tests
- Consistent prompt set
- All AI tools
- Screenshots as evidence
Competitive benchmarking
- Quarterly competitor mention analysis
- Share of voice calculation
Automated tracking
Web traffic
- GA4 with AI referrer filters
- UTM parameters for trackable links
Brand monitoring
- Google Alerts (limited for AI)
- Social listening tools
- Dedicated AI mention trackers (emerging market)
Reporting template
Monthly AI-search report
Executive summary
- Mention rate trend (vs previous month)
- Top 3 wins
- Top 3 improvement areas
Detail metrics
- Table with all KPIs
- Comparison vs previous period
- Comparison vs target
Competitor analysis
- Share of voice update
- New threats/opportunities
Action items
- Prioritized list of optimizations
- Resources needed
- Timeline
What you DON'T measure (and why)
AI tool analytics
Most AI tools don't offer publisher analytics. You can't see how often your content is used.
Exact click data
Many AI answers don't lead to clicks – the value is in the mention, not the click.
Real-time data
AI model updates are unpredictable. Measuring trends requires patience.
Best practices for measuring
1. Baseline first
Measure your current situation before optimizing. Without baseline, no progress.
2. Consistency
Always test the same prompts, in the same tools, at similar times.
3. Patience
AI models update irregularly. Give optimizations 4-8 weeks to show effect.
4. Correlation vs causation
Rising mentions after content updates are likely related, but not proven.
5. Business context
Link AI metrics to what matters: leads, revenue, brand awareness.
What can MatthCon do for you?
MatthCon sets up complete AI-search tracking for your brand. We deliver monthly reports with concrete optimizations and maintain your backlog. Start with a conversation or check our AI-search brand tracking service.