Organic traffic is falling across a lot of websites right now. If yours is one of them, the first instinct is to panic.
Don't. Much of the traffic AI search is taking was never going to turn into customers. The visitors who still arrive are often worth more than the ones you lost.
The real problem is that most companies still judge marketing by traffic. That was always the wrong scoreboard, and AI search has made it obvious.
The time to start measuring the right things is now, before a drop, not after it.
Why your traffic is falling
People now get answers without clicking. Pew Research found that Google users who saw an AI summary clicked a traditional search result in 8% of visits, compared with 15% when there was no summary. They clicked a link inside the summary itself just 1% of the time.
People also ask AI tools the questions they used to search for. Across a monthly sample of a million ChatGPT messages, looking up information rose from 16.7% in July 2024 to 20.2% in September 2026, while writing fell from 34.1% to 21.4%.
People now use ChatGPT to look things up
Share of ChatGPT consumer messages, from a monthly sample of one million messages
Source: OpenAI Signals data, via Search Engine Journal, October 2026.
Your buyers are using it that way at work too. Information seeking stayed around 11% to 12% of work messages for two years, then rose to 15.2% in September.
In the USA, requests to obtain information about goods or services rose from 1.7% to 2.9% of work messages.
Those answers rarely send a click. Ahrefs estimates ChatGPT handles about 12% of Google's search volume, yet Google sends 190 times more traffic to websites.
Why it's not as bad as it looks
Look at which pages are losing traffic. In most cases it's broad, informational content. Definitions, quick how-tos and general explainers that attract readers who were never going to buy.
AI tools answer those questions well, so those readers no longer need to click. Losing them hurts your traffic report far more than your revenue.
The visitors who do click through from AI tools are further along. They've already had the basic answer and want the detail, the product or the supplier.
Adobe found that AI referrals to US retail sites converted 31% better than other traffic over the 2025 holiday season.
Your brand can also be recommended without anyone visiting your site. When an AI tool names you as a supplier worth considering, that influence never shows up as a session. Traffic falls while influence can rise, and a traffic report alone can't tell the difference.
What to track instead
These are the measures I recommend to our clients.
| Instead of | Track | Where to find it |
|---|---|---|
| Total sessions | Revenue per visit, split by channel | GA4 conversion value and your CRM |
| Keyword rankings | AI share of voice, meaning how often AI answers name you against competitors | Ahrefs Brand Radar |
| Organic traffic | AI mentions and citations on your highest-value prompts, tracked separately | Ahrefs Brand Radar or a monthly prompt check |
| Pageviews per post | Pipeline influenced by content, the deals your content helped bring in | Your CRM, or a simple spreadsheet |
| Backlink counts | The third-party sources AI tools draw on in your category, such as Reddit, LinkedIn, YouTube and G2 | The cited sources in Ahrefs Brand Radar and your own prompt checks |
Add a "How did you hear about us?" field to your contact or demo form as well. Visits from AI tools often show up as direct traffic, and the field catches what analytics misses.
Be honest about the limits. Many people see your brand in an AI answer and never click, so the link between AI visibility and revenue is often a correlation, not proof. Track the trend over months, not the result of a single week.
For SaaS companies
SaaS buyers research heavily before they book a demo, and more of that research now happens in AI tools. The measures that matter most are the ones closest to pipeline.
Demo and trial requests by source, including AI referrals and the answers people give on your form
Pipeline influenced by content, especially comparison pages, case studies and integration pages
AI citations on comparison prompts, such as "best alternative to" a named competitor or "best tool for" your category
Mentions on review and community sites, such as G2, Capterra and Reddit, since AI tools draw heavily on them
We saw this in our content for field service SaaS, where practical, expert answers reached owners researching a problem long before they searched for software. For B2B, it's also worth knowing that ChatGPT ads don't reach most business buyers yet, so organic visibility in AI answers matters even more.
For ecommerce brands
For online stores, revenue per visit matters far more than the number of visits.
Revenue per visit, split by channel, so you can see which traffic actually buys
Orders from AI referrals, tracked as their own channel
Product visibility in AI shopping answers, now that tools like ChatGPT recommend Shopify products directly
Merchant feed health, since missing product data keeps you out of AI shopping results
Prune before you measure
Before you pay for a new tracking tool or an AI audit, count how many of your pages actually deserve to exist.
Most sites carry pages that are thin, outdated or competing with each other for the same topic. They dilute your site for Google and AI tools alike.
The approach that works is the same one that worked for Google ten years ago.
Cut or merge the weak pages. Redirect each one to the page that now answers the question best, so nothing with links or traffic is thrown away.
Build a smaller set of strong pages, each around a question real buyers ask that your site didn't answer well, or answered on several pages at once.
Get talked about elsewhere. Publish on LinkedIn, industry platforms and communities, all pointing back to those pages. When AI tools search the web, they often find third-party pages first, and those pages lead them back to you.
A smaller site with clear pages and plenty of mentions elsewhere tends to be picked far more often than a bigger, messier one.
We're doing exactly this on our own site. This year we've redirected more than 30 thin and outdated blog posts into stronger pages and rewritten the ones that matter, and our clicks are climbing again, up 28% week on week at the time of writing.
How we report it
Our monthly client reports bring this together in one place. GA4 tracks conversions and the AI channel, Search Console covers visibility and branded search and Ahrefs Brand Radar measures AI mentions and share of voice against competitors.
Looker Studio pulls it all into a single dashboard, so a client sees what's driving enquiries, not just what's driving visits.
You can start the same tracking yourself. It begins with GA4 set up properly, then AI citation tracking and a clear view of your content marketing ROI.
Our SEO, AI search optimisation and content writing teams work alongside PPC, social media and web design, so you can judge every channel by what it brings in.
-
Often because AI answers now satisfy questions that used to need a click. Pew Research found Google users clicked a search result in 8% of visits when an AI summary appeared, against 15% without one. Broad, informational pages lose the most. Check which pages are falling before you worry, because a drop on low-intent pages rarely affects enquiries or sales.
-
Yes, though not as your main measure of success. Traffic still shows whether people can find you, and a sudden drop can signal a technical problem. It no longer tells you whether marketing is working, because AI tools answer many questions without sending a click. Track it alongside enquiries, AI visibility and revenue by source.
-
Create a custom channel group in GA4 with a rule that matches AI sources such as chatgpt.com, perplexity.ai and gemini.google.com, and place it above Referral. Add a "How did you hear about us?" field to your forms too, because many AI-influenced visits arrive as direct traffic and analytics alone won't catch them.
-
Delete or merge the ones that are thin, outdated or competing with a better page on the same topic, and redirect each to the most relevant page so no links or traffic are lost. Keep and update the ones that still bring in the right readers. A smaller site with clear, strong pages tends to perform better in both Google and AI search.
