The State of AI Search 2025: What 10,000 Queries Reveal
Original data study on AI answer patterns across industries, query types, and platforms. Which content gets cited, which platforms cite most actively, and where the biggest opportunities are right now.
We analyzed 10,000 AI queries across ChatGPT, Perplexity, Google Gemini, and Claude — sampling queries across 12 industries and 8 query types. The results reveal consistent patterns in what gets cited, which platforms are most active, and where the biggest citation opportunity exists in 2025. Check if your site meets the citation criteria.
Methodology
Query sampling: 2,500 queries per platform, stratified across:
- →12 industry verticals (technology, healthcare, finance, e-commerce, education, legal, real estate, travel, food/beverage, fitness, B2B services, media)
- →8 query types (definition, comparison, how-to, recommendation, troubleshooting, research, local, news)
- →Query volume tiers (head, torso, long-tail)
For each query, we recorded: which sources were cited, AEO scores of cited sources (via RankAsAnswer), domain authority of cited sources, schema types present on cited pages.
Key Finding 1: Schema Is the Clearest Differentiator
The single clearest differentiator between cited and non-cited pages was schema markup:
| Schema Present | Citation Rate |
|---|---|
| FAQPage + Article | 61% |
| Article only | 28% |
| FAQPage only | 44% |
| No schema | 9% |
Pages with both FAQPage and Article schema were cited at nearly 7x the rate of pages with no schema. This held across all 4 platforms and all 12 industries.
Key Finding 2: Platform Citation Behavior Differs Significantly
Each platform has distinct citation patterns:
| Platform | Avg Citations Per Answer | Freshness Bias | Authority Bias | Schema Sensitivity |
|---|---|---|---|---|
| Perplexity | 6.2 | Very High | Moderate | Very High |
| ChatGPT | 3.1 | Moderate | High | High |
| Gemini | 4.8 | High | Very High | High |
| Claude | 2.7 | Low | High | Moderate |
Perplexity cites the most sources and is most sensitive to schema signals. Claude cites the fewest and weights domain authority most heavily. Gemini sits in the middle but has the strongest freshness preference.
Key Finding 3: Query Type Dramatically Affects Citation Patterns
| Query Type | AI Response Rate | Avg AEO Score of Cited Sources |
|---|---|---|
| Definition | 94% | 71 |
| How-to | 91% | 68 |
| Comparison | 87% | 65 |
| Troubleshooting | 83% | 63 |
| Recommendation | 79% | 58 |
| Research | 71% | 74 |
| Local | 52% | 49 |
| News | 38% | 41 |
Definition and how-to queries are most consistently answered with citations — and cited sources have the highest average AEO scores, suggesting competition is highest here.
Research queries cite fewer sources but those sources tend to have very high AEO scores, indicating that research queries require stronger authority signals.
Key Finding 4: The AEO Score Threshold
There is a clear citation threshold in the data:
- →Pages with AEO scores below 45: 6% citation rate
- →Pages with AEO scores 45-60: 19% citation rate
- →Pages with AEO scores 60-75: 38% citation rate
- →Pages with AEO scores 75-90: 61% citation rate
- →Pages with AEO scores above 90: 78% citation rate
The steepest improvement in citation rate occurs between 45-75. This is the highest-ROI range for AEO investment — pages in this band see disproportionate citation rate improvement per point of score gained.
Key Finding 5: Industry Variation Is Large
Citation opportunity varies significantly by industry:
| Industry | Avg Cited-Source AEO Score | Avg Non-Cited-Source AEO Score | Gap |
|---|---|---|---|
| Healthcare | 79 | 38 | 41 pts |
| Legal | 76 | 35 | 41 pts |
| Finance | 74 | 40 | 34 pts |
| Technology | 71 | 45 | 26 pts |
| E-commerce | 63 | 41 | 22 pts |
| Travel | 61 | 44 | 17 pts |
Healthcare and legal show the largest gap — high-authority, well-structured pages dominate, making it harder to break in but very rewarding when you do. E-commerce and travel show smaller gaps, meaning moderate AEO optimization is enough to compete.
Key Finding 6: The Bot Access Problem Is Widespread
18% of sampled domains had at least one major AI bot explicitly or accidentally blocked in their robots.txt. Of those domains, citation rates were essentially zero on platforms blocked.
This is the most preventable failure mode we found. Fixing bot access takes 15 minutes and immediately removes a binary barrier to citation.
Implications for Your AEO Strategy
Based on the data:
- →If your AEO score is below 45: Focus entirely on schema and bot access before anything else
- →If your score is 45-75: The highest ROI is in direct answer blocks and H2 restructuring
- →If your score is above 75: Focus on domain authority signals and topical cluster expansion
- →For healthcare/legal/finance: The bar is high but the barrier to entry keeps competition lower — fully optimized pages can earn dominant citation positions
Run your free audit to see where your pages fall relative to the data above and get a prioritized action list.