Is SEO Traffic Declining Because of ChatGPT or AI Overviews?

Lots of people in SEO are talking about how organic traffic is down in 2025. Some marketers are even reporting “massive” (40%+) drops in traffic and declaring that SEO is dead (it’s not, keep reading). 

We’re not seeing such extreme declines in organic traffic across our 20+ clients, but yes, there’s been some drop.

So in this post, we want to report on that data and answer two key questions:

• What’s actually causing this decline — AI Overviews, ChatGPT, or something else?
• Should marketing teams be worried? And if so, what should they do? 

The Data: Traffic Is Down but Rankings and Conversions Are Growing

We’re seeing a consistent pattern across several clients: more rankings but less traffic. 

Here’s a typical example from one client’s Ahrefs rankings in October 2023, 2024, and 2025:

And here’s that same data shown in a bar graph, so you can easily see the growth in rankings:

But look at this client’s traffic, measured by clicks in Google Search Console (displayed via SEO Gets). The blue line represents clicks:

There’s a slight decline in clicks from mid-2024 to October 2025. 

Notably, impressions (purple line) are trending up. That means our organic listings on the Google SERP are showing up more often now than before, but earning fewer clicks (blue line). The green line, click-through rate (CTR), confirms this with a sharp drop around March 2025. 

We’ll discuss why that date is important below, but this is the trend we’re seeing across multiple clients. 

For instance, here are two more clients’ Search Console graphs showing impressions (purple) increasing while clicks (blue) decrease: 

Finally, what’s interesting is that despite the drop in traffic, clients are seeing steady growth in conversions. 

Here’s data on free-trial conversions attributed to our content for a SaaS client from early 2025 to now:

What’s going on? 

How are businesses seeing record-high rankings, declining traffic, yet steady (or even growing) conversions?

In all our years running Grow and Convert, rankings, traffic, and conversions have always grown together — when rankings and traffic were down, conversions were also down.

But now we’re seeing two metrics rising (rankings and conversions) while traffic falls.

So, why is that?

Why We Think It’s AI Overviews, Not ChatGPT, Causing the Traffic Decrease 

We believe the decline in clicks and site traffic is due to Google’s AI Overviews having more prominence in search, not from users moving off Google for ChatGPT. 

Why ChatGPT Is Likely Not Causing the Traffic Drop

First, in the Search Console data above, impressions are holding steady — and for many clients, they’re actually growing. If the organic traffic decline was caused by users shifting from Google to ChatGPT, we’d expect to see overall impressions drop too. We don’t.

Second, in our clients’ analytics data, traffic coming from ChatGPT is less than 5% of what they get from Google Organic. 

Here’s that same data as a bar graph, to make the difference visually clear:

Ahrefs created a site called ChatGPT vs. Google, and they’re largely seeing the same thing — traffic from ChatGPT remains minimal compared to Google, Bing, and other traditional search engines.

A study from Semrush also shows that ChatGPT is now used as one step in the search process. People might start their search in ChatGPT, but they still turn to Google to complete their searches. So ChatGPT usage alone doesn’t explain the traffic decline.

You can read the full study here.

So if ChatGPT isn’t the cause, what about Google’s AI Overviews?

Why Google AI Overviews Is Likely Causing the Traffic Drop

Notice in the graphs above there’s a sharp drop in clicks (blue line) and CTR (green line) around March 2025. 

I was curious about that timing, so I asked Google about the rollout of AI overviews, and March 2025 seems to be when they rolled it out more broadly:

This strongly suggests that the drop in clicks (and traffic) starting around March 2025 is related to AI Overviews. 

The next question is how and why AI Overviews can reduce traffic without hurting impressions, even as rankings continue to rise.

Why AI Overviews Cause a Decline in Organic Traffic Even as Rankings and Conversions Grow

First, for top-of-funnel queries you used to rank for, AI overviews now answer the user’s question directly — so they no longer need to click into your article. That’s likely the primary reason brands are reporting drops in organic traffic. 

It also explains why we at Grow & Convert are seeing smaller declines in traffic (10%–20%) compared to others reporting 40%–50%. We focus on ranking clients for bottom-of-funnel search terms, meaning queries where users are looking for a product or service like theirs. For those terms, even if an SEO article doesn’t get as much traffic as before, the brand still gets mentioned in the AI overview itself. 

For example, here’s the AI Overview for the term “best call center quality assurance software,” where our client Level AI is mentioned in the AI Overview, and a page on their site is one of the cited sources on the right:

A user looking for software like this isn’t really incentivized to click on the articles on the right — the AI Overview already summarizes which brands are popular. That’s the simple, obvious explanation for the drop in traffic. But there’s also something more subtle happening here. 

Specifically, notice how Level AI is mentioned multiple times in the AI Overview. That means several different user paths could unfold.

Scenario 1: New Tab

The user might just open a new tab, Google “Level AI,” click into the homepage, and request a demo or convert. That conversion would appear as organic traffic to the homepage — not the article — even though the article was instrumental in getting Level AI featured in the AI Overview in the first place. As we’ll discuss below, that creates a major attribution problem. 

Note that if the user’s eyes go to the second bolded “Level AI” mention under “top call center QA software”, there’s no link, which makes it even more likely they’ll just open a new tab and search “Level AI.” 

Scenario 2: Dotted Line Link

Notice how the first mention of Level AI in the AI Overview introduction has a thin dotted underline. When you click that link, it triggers a “disguised” branded search. 

It looks like a normal branded search, but if you check the URL and the “showing results for” line, you’ll see it actually searches for “Level AI call center quality assurance software.” 

As a result, the Level AI homepage isn’t the top result — instead their Quality Assurance Software page appears first, followed by a related blog post, and then the homepage. And all three of those organic links are preceded by a sponsored spot because, well, Google’s gotta make money. 

When this happens, traffic from that search term can show up as any of the following: 

• Branded paid (if the user clicks any of the sponsored links)
• Organic to the feature page 
• Organic to the blog post
• Organic to the homepage

It’s confusing! 

What this means is that traffic that used to be easy to attribute (one search term corresponding to one URL) is now spread across a mishmash of possibilities: users might go to the homepage, a feature page, a paid ad, or a blog post. 

Which brings us to the real consequence for marketers: attribution is getting even harder. 

Attribution Needs to Change Because You’re Not Measuring the Full Benefit of SEO Anymore

SEO attribution has always been tricky, and since the start of our agency, we’ve aimed to fix that. We’ve made a deliberate effort to track leads all the way through the funnel so we can see which blog posts convert best.

We’ve always said our attribution represents a lower-limit estimate of the actual number of conversions our services generate, mainly because there are a number of scenarios where you can’t measure traffic all the way through conversion.

Specifically:

• If a user comes in through SEO but has cookies blocked, your SEO/content efforts won’t be attributed to that lead.
  
• If someone discovers your brand through SEO but switches devices before converting, that conversion won’t be attributed to SEO.
  
• If someone finds your brand through SEO and tells a coworker or friend who later searches for it separately, that lead won’t be attributed to SEO.

Now, with the rise of AI overviews, there’s yet another attribution challenge. If people discover your company through an AI overview, what was previously attributed to organic search now often appears as direct or branded traffic.

Why?

Because, as shown above, AI overviews display company names as clickable links that open branded searches. Many users might also open a new tab and type your domain directly or search for your brand manually. 

In other words, SEO is still driving visibility — it’s what gets your brand mentioned in the AI overview — but the resulting increase in branded and direct conversions isn’t being credited to SEO, even though it originates there.

What’s the Best Way to Account for This Change in SEO?

We’re advising clients to add a “How did you hear about us?” field to their lead forms. This lets prospects self-report whether they came from Google, AI, referrals, or another channel — providing another way to attribute leads. We’re also encouraging sales teams to ask the same question during qualification calls.

In addition to tracking organic traffic to our articles, we’ve started measuring organic homepage traffic as well. We’ve also created additional reports that identify referral sources from LLM platforms to see how much traffic is coming from LLMs.

Overall, we’re resetting expectations around traditional organic leads and traffic metrics and looking at organic leads more holistically — across both homepage and article traffic.