Yes, AI search still sends traffic, but less evenly than classic search. Quick factual questions often get answered on the results page with no click, while research-heavy, local and buying-intent searches still produce visits. The catch is measurement: Google doesn't separate AI Overview clicks in Search Console, so you have to piece the picture together.
That combination, real but smaller traffic you can't see clearly, is why so many business owners feel uneasy about AI search. The answer isn't panic or denial. It's setting up measurement that shows which queries still bring visitors and, more importantly, which visitors still become enquiries. For the strategy side, start with our guide to AEO, GEO and ranking in AI Overviews and ChatGPT.
Why is AI search traffic so hard to see?
Because the main reporting tools weren't built to label it. Google folds AI features into its normal search data, and many clicks from AI apps arrive without telling your analytics where they came from.
Three blind spots cause most of the confusion:
- Search Console blends it in. Clicks and impressions from AI Overviews and AI Mode are counted inside the standard Web search performance report. There's no filter that shows "AI Overview clicks only".
- Referrers go missing. When someone taps a link inside a mobile AI app, the visit sometimes lands in analytics as "Direct" because no referrer was passed.
- Influence without a click. Someone may read your brand name in an AI answer, then search for you by name a day later. Analytics credits branded search, not the AI answer that started it.
So "my traffic is down" and "AI search doesn't work for me" are not the same statement. You need to look at which queries changed and whether enquiries changed with them.
Which searches still send clicks, and which don't?
Searches that need a decision, a local provider or depth still send clicks; searches that need a single fact send fewer. This is a pattern to test against your own data, not a law, but it's a sensible starting point.
| Type of search | Example | Likely click behaviour with AI answers | What to do |
|---|---|---|---|
| Quick definition or fact | "what is a canonical tag" | Often answered without a click | Aim for the citation and brand exposure, not the visit |
| How-to with several steps | "how to set up Meta lead ads" | Some clicks, especially from people doing the task | Offer screenshots, templates, detail the summary can't hold |
| Comparison or "best" | "Shopify vs WooCommerce for a small store" | Clicks from people close to deciding | Clear tables, honest trade-offs, pricing context |
| Local service | "wedding photographer in Patna" | Strong click and call intent, often via Maps | Google Business Profile, reviews, fast mobile pages |
| Pricing and quotes | "custom CRM cost India" | Clicks from serious buyers wanting specifics | Real ranges and what drives them |
| Branded | "GRWW Ops Console" | High click-through to your site | Make your own pages the best source on you |
If most of your traffic loss sits in the first row, you've probably lost visitors who were never going to become customers. If it sits in the bottom three rows, that's worth investigating urgently.
How do you estimate AI Overview impact in Search Console?
Compare impressions against clicks query by query over time, and look for rising impressions with falling click-through rate on informational searches. That pattern is the clearest indirect sign that answers are being satisfied on the results page.
A practical method:
- Export 16 months of query data from the Performance report so you can compare like-for-like periods.
- Group queries by intent. Tag each one as informational, comparison, local, pricing or branded. A spreadsheet with a simple formula on keywords like "what", "how", "vs", "cost" and "near me" gets you most of the way.
- Compare click-through rate by group, not just total clicks. Overall averages hide what's happening.
- Check the results yourself. For your top losing queries, search them and note whether an AI Overview appears and whether you're cited in it.
- Annotate changes. Note Google updates and your own site changes so you don't blame AI for a problem caused by a redesign.
Our walkthrough on optimising for Google AI Mode explains how to improve your chance of being one of the cited links once you know which queries matter.
How do you track visits from ChatGPT, Perplexity and Gemini?
Create a custom channel in Google Analytics 4 that groups AI referrers together, so they stop hiding inside "Referral". It takes about 15 minutes.
Steps in GA4:
- Go to Admin, then Data display, then Channel groups, and copy the default channel group.
- Add a new channel called "AI assistants".
- Set the condition to session source matching a regular expression such as
chatgpt\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai. - Move the new channel above "Referral" in the list, because GA4 assigns sessions to the first matching rule.
- Save, then use the channel group in your acquisition reports.
ChatGPT also adds a utm_source=chatgpt.com tag to many outbound links, which makes those visits easier to spot. Review the list every few months, because new assistants and domains appear.
Remember the limits: this only captures visits that pass a referrer. Treat the number as a floor, not a total.
What should you measure instead of raw traffic?
Measure enquiries and where they came from, because a smaller number of better-informed visitors can be worth more than lots of casual ones. Someone who arrives after reading an AI summary often already knows what they want.
Metrics worth tracking monthly:
- Enquiries by channel, including the new AI assistants channel.
- Conversion rate of AI-referred sessions compared with organic search.
- Branded search impressions in Search Console, a rough signal of growing awareness.
- Citation share: how often you appear in AI answers for a fixed set of questions, covered in our guide to tracking your brand's visibility in AI answers.
- Self-reported source: a "How did you hear about us?" field with "ChatGPT or another AI assistant" as an option.
That last one sounds old-fashioned, but it catches the influence analytics misses. In our own Ops Console, the lead centre pulls website form and Meta lead-ad enquiries into one place automatically, so the source field and the enquiry sit side by side instead of in separate tools. Whatever system you use, the principle is the same: connect the visit to the lead. Our guide to tracking whether your website is getting you leads shows how to set that up.
Should you still invest in content if AI answers reduce clicks?
Yes, but shift effort toward content that earns citations, captures decision-stage visitors and supports your brand. Generic informational posts written purely for traffic are the weakest investment now.
Practical priorities:
- Strengthen pages close to revenue: services, pricing, comparisons, case studies.
- Publish what summaries can't replace: original data, detailed processes, tools, templates.
- Keep informational content where it builds trust and earns citations, even if clicks are lower.
- Make every visit count: fast pages, clear calls to action, easy WhatsApp or call options on mobile.
None of this needs a big team: a monthly hour with Search Console, GA4 and your enquiry list is enough to see which content still pulls its weight.
Frequently asked questions
Can I see AI Overview clicks separately in Google Search Console?
Not at the time of writing. AI Overview and AI Mode clicks and impressions are included in the standard Web search performance data without a separate filter. You can infer impact by comparing click-through rates by query type over time.
Why does ChatGPT traffic show up as direct traffic?
Some AI apps, especially on mobile, don't pass a referrer when a user taps a link. Analytics then has no source to record and labels the visit as Direct. A self-reported source field on your enquiry form helps fill that gap.
Is a drop in organic traffic always caused by AI Overviews?
No. Site changes, technical errors, algorithm updates, seasonality and competitors can all cause drops. Check your own changes and the specific queries affected before concluding AI is responsible.
Are visitors from AI assistants better leads?
They can be, because they've often done research before clicking, but it varies by business. Compare the conversion rate of your AI assistants channel with organic search over a few months rather than assuming either way.
How often should I review AI search traffic?
Monthly is enough for most small businesses. Look at enquiries by channel, AI-referred sessions and a fixed set of citation checks, then make one or two content decisions based on what you see.
Want measurement that shows what's really working?
We set up GA4 channels, Search Console reporting and lead tracking that connects visits to enquiries, so you can make decisions on real data. If you'd like it done properly, get a fixed-price quote.