At a glance
Sessions referred from generative AI platforms (ChatGPT, Perplexity, Claude, Gemini, Copilot). AI search is the fastest-growing traffic channel in ecommerce, growing roughly 10x year-over-year as of 2026. The card surfaces presence and volume so merchants can detect when AI agents are recommending their store. Even a few hundred sessions per month from this channel is a positive signal that brand and product copy are AI-discoverable; zero suggests catalogue gaps (missing descriptions, broken structured data) blocking AI surfacing.
| What it counts | Sessions in GA4 with session_source matching the known AI platform domains (chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, copilot.microsoft.com, you.com, and emerging platforms). Excludes traditional search and direct traffic. |
| Sample type | GA4 session-level data, refreshed on the standard data refresh. |
| Why AI traffic matters | (1) Net-new channel growth: AI search is currently 1-3% of ecommerce sessions but growing 10x annually; presence today positions for outsized share tomorrow. (2) High-intent visitor profile: AI users typically arrive after a specific question (e.g., “best garden hoses for hard water”), pre-qualified for purchase. (3) Channel dependency signal: stores with zero AI traffic have content/SEO gaps that exclude them from AI candidate sets. (4) Brand discovery proxy: AI recommendations are based on training-data + live-browse content; high AI traffic means the brand is well-represented in both. |
| Reading the value | (1) 0 sessions: catalogue is invisible to AI. Investigate structured data, descriptions, and brand authority gaps. (2) 1-100 sessions: emerging presence; track growth. (3) 100-1,000 sessions: meaningful contribution; optimise. (4) Above 1,000: significant channel; treat as a primary marketing-priority channel. (5) Cross-reference ai_conversion (conversion rate of AI visitors), ai_revenue (revenue contribution), ai_engagement (quality of interactions). |
| Currency | absolute count (sessions). |
| Time window | 30D vsP. |
| Alert trigger | ai_traffic = 0 (BAD threshold). Any presence is healthy; total absence indicates a discoverability gap. |
| Sentiment key | ai_traffic (HIGHER_IS_BETTER; GOOD ≥ 1, BAD = 0). |
| Roles | owner, marketing |
Calculation
known_ai_domains includes (non-exhaustive, updated as new platforms emerge):
- chatgpt.com (OpenAI ChatGPT browsing)
- perplexity.ai (Perplexity)
- claude.ai (Anthropic Claude)
- gemini.google.com (Google Gemini)
- copilot.microsoft.com (Microsoft Copilot)
- you.com (You.com search)
- bing.com via Bing Chat referrer
- search.brave.com (Brave Search AI summaries)
Worked example
A UK-based ecommerce merchant on BC, AI Traffic reading on Wednesday 15 May 26.| AI Platform | Sessions (30D) | % of AI traffic | vs prior period | AI conversion rate | Notes |
|---|---|---|---|---|---|
| ChatGPT (chatgpt.com) | 184 | 56% | +42% | 3.2% | Largest AI source |
| Perplexity (perplexity.ai) | 78 | 24% | +28% | 4.1% | Highest-converting AI source |
| Google Gemini | 38 | 12% | +18% | 2.4% | Growing |
| Microsoft Copilot | 18 | 5% | +60% | 1.8% | Edge integration driving growth |
| Claude (claude.ai) | 9 | 3% | +200% | 5.5% | Smallest but highest CR |
| Total AI traffic | 327 | 100% | +38% vsP | 3.4% (avg) | Performing Well |
- 327 sessions / 30 days from AI platforms. Roughly 11 sessions/day. Small in absolute terms vs paid (typically 1,000+/day) but growing 38% period-over-period, extrapolating, this becomes 800+ AI sessions/month within 6 months and 3,000+ within 12 months at current growth.
- AI traffic converts at 3.4% on average, above the merchant’s site-wide conversion rate of 1.94%. This is the consistent pattern: AI users arrive having asked a specific question, are pre-qualified, and are closer to purchase than typical organic visitors. AI traffic is small but quality-dense.
-
Platform mix:
- ChatGPT dominates (56%): highest absolute volume because of broadest user base.
- Perplexity (24%): punches above weight because it’s purpose-built as a search alternative.
- Gemini (12%): growing as Google integrates AI summaries into core search.
- Copilot + Claude (8% combined): small but fast-growing as users adopt non-Google AI.
- Why ChatGPT and Perplexity send the most. Both have explicit “browse the web” or “show sources” features that link out to merchant sites. Models without explicit source-attribution (older Claude versions, base GPT) don’t refer traffic because they answer in-line without linking out. As all AI platforms add source attribution, AI traffic share should grow.
-
What “zero AI traffic” looks like and why it’s bad. A store showing 0 in this card has either: (a) thin product copy that AI can’t extract product knowledge from, (b) broken structured data (Product schema missing or invalid), (c) low brand authority such that AI doesn’t surface the merchant when generating product recommendations, (d)
noindexor robots blocks preventing AI crawlers from accessing the site. Each is fixable, and each fix opens the channel. -
Recommended actions to grow AI traffic:
- Audit product copy quality: rich, specific descriptions including materials, use cases, sizing details. Generic copy doesn’t get cited.
- Validate structured data: Product schema must include
name,description,sku,offers.price,aggregateRating,review. Use Google Rich Results Test. - Brand authority: AI training data prefers authoritative sources. Build PR mentions, Wikipedia presence, structured data on About pages.
- Don’t block AI crawlers: ensure
robots.txtdoesn’t block GPTBot, ClaudeBot, PerplexityBot, etc. (Some merchants block these to avoid training-data scraping; the trade-off is exclusion from AI search.) - Cross-reference:
missing_desc,missing_seo(catalogue health gaps that block AI discovery).
- Read the count. Above zero is healthy; zero is a discoverability gap.
- Decompose by platform. Identify which AI platforms are sending traffic.
- Cross-reference conversion rate (
ai_conversion) to confirm quality. - For zero or growing: audit content + schema + crawler access for AI-source eligibility.
| Time horizon | Action |
|---|---|
| First 1 hour | Read count. Decompose by platform. |
| First day | Cross-reference conversion rate. |
| First week | If zero, audit structured data + content + crawler access. |
| Quarterly | Track growth trajectory; AI is the fastest-growing channel. |
Sibling cards merchants should reference together
| Card | Why merchants reach for it |
|---|---|
ai_conversion | AI conversion rate; quality measure. |
ai_revenue | AI revenue contribution. |
ai_engagement | AI engagement rate; depth of interaction. |
ai_traffic_overview | Detailed AI traffic breakdown. |
ai_traffic_trend | AI traffic over time. |
ai_vs_search | AI vs traditional search comparison. |
gsc_clicks_trend | Organic search clicks; complementary channel. |
missing_desc | Catalogue copy quality; blocks AI discovery. |
Reconciling against the vendor’s own dashboard
Where to look in GA4: Reports → Acquisition → Traffic acquisition; filter by source containingchatgpt, perplexity, claude, gemini, copilot, you.com, etc.
Why our number may differ:
| Reason | Direction | What to do |
|---|---|---|
| AI domain list. Vortex IQ maintains a curated list of AI platforms; GA4 doesn’t categorise them automatically. | Vortex IQ may include more platforms | Confirm the list. |
| Bot filter. GA4 default bot filter may catch some AI crawler traffic; this card excludes bot-flagged sessions. | Match | n/a. |
| Period boundary. GA4 default views may differ. | Variable | Match periods. |
| Cross-domain tracking. AI platforms may strip referrer; some sessions land as direct rather than AI. | Vortex IQ may understate | Use UTM-tagged links from AI-summary platforms when possible. |
gsc_clicks_trend (organic search) and direct traffic to see the full discovery channel mix.
Quick rule: confirm the AI domain list and check for any new platforms not yet in the curated list.
Known limitations / merchant FAQs
Q: Our AI traffic is 0. What’s wrong? Three likely causes: (a) product descriptions are thin, generic, or missing, AI can’t extract product knowledge from empty PDPs; (b) structured data is broken or missing, Product schema is what AI reads first; (c) AI crawler bots (GPTBot, ClaudeBot, PerplexityBot) are blocked in robots.txt. Audit each in order; the fix is straightforward and opens the channel. Q: AI traffic seems too small to care about. Should I ignore it? Not for long. AI search is currently 1-3% of ecommerce sessions but growing 10x year-over-year. Stores that establish AI presence now will compound advantage; stores that wait until AI is 15-20% of traffic will face the same SEO catch-up problem stores experienced with mobile in 2014-2018. The investment to get AI-discoverable is small (content + schema), the payoff compounds. Q: How do I know if my brand is being mentioned by AI agents (without traffic)? Periodically prompt-test: ask ChatGPT, Perplexity, Claude, Gemini “what’s the best [your category] store in [your country]?” If your brand appears in the answer, you’re in the candidate set. If not, the answer reveals which competitors are; reverse-engineer their content/schema/authority advantages. Q: AI agents block my robots.txt, should I unblock them? Default to yes for ecommerce. The trade-off: blocking prevents your content from being used as training data (which some merchants prefer for IP reasons), but also excludes you from AI search results. For most ecommerce, the discoverability benefit dwarfs the training-data concern. Q: My AI conversion rate is 5%+ but volume is tiny. Should I scale it? Not directly via “spending”, AI traffic isn’t paid. But the levers that grow AI traffic (rich product copy, structured data, brand authority, PR coverage) all also lift organic search and direct conversion. Investing in AI-discoverability has compound benefits across all organic channels. Q: How does this differ fromgsc_clicks_trend (organic clicks)?
gsc_clicks_trend measures clicks from Google’s traditional 10-blue-link search results. This card measures sessions from generative AI conversations. They overlap when Google’s AI Overviews link to your site (counted as Gemini), but most AI traffic comes from non-Google AI platforms (ChatGPT, Perplexity, Claude) which are entirely separate.
Q: Why is the threshold at “any > 0” rather than something more aggressive?
Because AI traffic is a brand-new channel. Setting a higher threshold (say 100 sessions) would mark 70%+ of merchants as failing on a still-emerging channel. The current threshold (any > 0) signals “are you discoverable at all”, the right question for 2025-26. Expect this threshold to tighten over time as the channel matures.
Q: My AI traffic dropped 50% this period. What happened?
Common causes: (a) AI platform updated its source-attribution logic and now references your site less; (b) your structured data validation broke after a recent release; (c) a competitor improved their content and is being preferred; (d) seasonal: query patterns shift. Investigate via the per-platform breakdown to see whether the drop is concentrated on one platform (likely a structured-data or bot-block issue) or distributed (likely competitive or seasonal).