State of AI Search 2025: The Shift to Answer Engines

Search behavior is changing. Users are moving from "searching" on Google to "asking" on AI Answer Engines.

At ViaMetric, we analyzed traffic patterns across our network to understand this shift. This report serves as a benchmark for the State of AI Search in 2025.

Key Findings (Executive Summary)

• AI Referral Growth: Traffic from AI sources (ChatGPT, Perplexity, Claude, Gemini) has grown by 300% Year-over-Year.
• The "Zero-Click" Reality: 65% of AI interactions do not result in a click-through, prioritizing Share of Voice over traditional CTR.
• Perplexity's Conversion Rate: Visitors from Perplexity span 2.5x longer on site than visitors from Google Organic.

1. Top Traffic Sources by Category

While Google still dominates volume, AI sources dominate intent.

Source Channel	Traffic Share	YoY Growth	Avg. Session Duration
Google Search	68%	-12%	1:45
Direct / Dark	18%	+5%	2:10
Perplexity AI	8%	+450%	4:20
ChatGPT	4%	+210%	3:15
Other AI (Claude/Gemini)	2%	+150%	2:40

Analyst Note: Perplexity users are "high-intent" researchers. They have already read the summary; if they click, they are ready to buy.

2. Brand Citation Frequency

We tracked how often B2B SaaS brands are mentioned in "Best X for Y" queries.

• Top 10% of Brands: Mentioned in 85% of relevant AI answers.
• Average Brand: Mentioned in only 12% of relevant AI answers.
• Optimization Impact: Brands using JSON-LD Schema appeared 50% more often than those without.

3. The "Hallucination Gap"

How accurate are AI mentions?

AI Engine	Accuracy Rate	Hallucination Rate
GPT-4o	96%	4%
Perplexity (Sonar)	98%	2%
Google Gemini	92%	8%

Methodology

This data is based on an aggregated analysis of anonymized traffic data processed by the ViaMetric platform between Jan 2024 and Jan 2025. "AI Referral" is defined as traffic with a Referrer or User-Agent matching known AI crawlers and chat interfaces.

Cite this report as: "ViaMetric State of AI Search 2025 Report".