AI tools are changing how people find information. Google’s AI Overviews, ChatGPT, Perplexity, and similar platforms now answer questions directly. Users read those answers and often stop there. Website owners see fewer clicks. Marketing teams see traffic reports trending down. The concern is real.
But the question worth asking is this: is traffic actually dying, or is it moving somewhere new?
The answer matters because it changes what you do next. If traffic is dying, you defend. If it is moving, you follow it.
The Traffic Drop: How AI Answers Are Reducing Traditional Clicks
Google’s AI Overviews launched widely in 2024 and immediately changed click behavior. Research from SE Ranking showed that AI Overviews appear for roughly 47% of all queries. When they appear, click-through rates on organic results drop significantly.
Ahrefs published data showing that informational queries, the ones that used to drive strong blog traffic, now show the steepest click decline. Users get a summarized answer at the top of the page and move on without clicking anything.
HubSpot reported in 2024 that its organic traffic dropped by roughly 80% on some informational pages after AI Overviews became widely visible. This was not a minor shift. Entire content categories lost most of their traffic within months.
The pattern is consistent. AI answers hit hardest on three content types:
- Definition-style content (“What is X”)
- How-to content with short, factual answers
- Comparison content where a simple summary satisfies the query
If your site is built traffic on these content types, the drop is real. But this is not the full picture.
Transactional queries, local searches, and queries requiring personal judgment still drive clicks. Someone searching “best accountant in Bishkek” or “which CRM should I use for a 20-person team” still clicks through to find a real answer. AI gives a general response but not a trusted, specific recommendation. That gap still belongs to you.
The traffic drop is selective. It targets low-effort content that AI can replicate instantly. That content was always fragile. AI just exposed the weakness faster than most site owners expected.
The Behavior Shift: Why Users Get Answers Before Visiting Websites
User behavior has shifted at the intent level. People now expect an answer before they click. This is not entirely new. Google’s featured snippets did this years ago. AI Overviews and tools like Perplexity push it further and faster.
Perplexity grew from roughly 10 million to over 100 million monthly queries between early 2023 and late 2024. Users go there specifically to get synthesized answers without visiting multiple sites. The site visit itself has been replaced by the answer, at least for certain query types.
This shift changes what discoverability means in practice. Being visible no longer requires being clicked first. Brands and publications now appear inside AI-generated answers. That appearance builds familiarity. The click comes later, sometimes much later, when the user is ready to take action.
This is where traffic measurement gets complicated. Standard analytics tools measure sessions and pageviews. They do not measure how many times your brand name appeared in a ChatGPT response or a Perplexity summary. That exposure is real, but it is invisible to traditional tracking.
Some teams use a traffic bot or crawl simulation tool to understand how AI agents and search crawlers interact with their content structure. This helps identify whether pages are being indexed and referenced correctly by AI systems, which increasingly determines whether your content gets surfaced in AI-generated answers at all.
The behavior shift you need to understand is this: users now make two separate decisions. First, they decide whether your brand is credible based on AI-mediated exposure. Second, they decide whether to visit your site. Both decisions matter. You used to only compete at the second stage. Now you compete at both.
The sites winning in this environment give AI systems enough structured, specific, well-sourced content to reference. When an AI tool cites your data or references your framework, you exist in the user’s awareness before they ever click a link.
The New Referral Map: Where Discovery Moves in an AI-Mediated Web
Traffic is not disappearing. It is redistributing. The referral sources that matter in 2025 look different from those that mattered in 2020. Search is still large, but it now competes with AI chat interfaces, newsletter platforms, community sites, and video.
Here is where traffic is actually moving.
Newsletters and owned audiences. Morning Brew grew to over 4 million subscribers by focusing on content people want delivered directly. Traffic from email bypasses search entirely and is immune to algorithm changes. Brands with strong email lists saw more stability through the 2024 AI disruption than those relying on organic search alone.
Reddit and community platforms. Google’s 2024 Helpful Content updates pushed Reddit results higher for many informational queries. Reddit now appears prominently for product reviews, comparisons, and real user experiences. If your brand is part of those conversations, you gain indirect visibility. If you are absent, someone else’s opinion fills the space.
YouTube. Video content sits outside most AI answer formats. Users still watch tutorials, reviews, and demonstrations. Brands like Ahrefs, Semrush, and Backlinko built substantial YouTube audiences that now serve as stable traffic sources independent of search algorithm changes. A well-structured YouTube channel functions as a second search engine presence for your brand.
Direct traffic and branded search. When users encounter your brand through AI exposure, newsletters, or community discussions, they start searching directly for you. Direct traffic and branded search volume signal growing trust. These are harder to build than capturing a generic Google click, but they are far more durable.
AI referral traffic. This channel is still emerging. Perplexity, ChatGPT with browsing enabled, and similar tools generate direct referral traffic when they link to source content. It is a small share for most sites today, but it is growing. Sites that appear as cited sources in AI outputs are building a new referral channel that did not exist three years ago.
The referral map has expanded. Discovery now happens across more surfaces. Sites losing traffic are those that built everything on one source, usually Google organic search. Sites maintaining or growing total reach treat discovery as a multi-channel process from the start.
The Real Verdict: AI Isn’t Killing Traffic, It’s Forcing Brands to Earn It Differently
Traffic is shifting, not dying. The data supports this.
Sistrix analyzed over 800,000 domains after Google’s AI Overview rollout and found that high-authority, well-structured sites held their traffic better than thin content sites. The correlation between content depth and traffic stability was consistent across categories.
The New York Times, The Wirecutter, and similar editorial brands reported steady or growing direct traffic through the same period that hit informational blogs hard. Their content is opinionated, experience-based, and specific in ways that AI cannot accurately replicate.
What separates brands that keep traffic from those that lose it comes down to three things.
Original data and research. AI tools cite sources. They reference sites that publish surveys, studies, and original findings. If your content is a rewrite of what already exists, AI can produce the same output without sending anyone to your site. If your content contains something AI cannot generate independently, your site becomes a source rather than a competitor.
Clear brand positioning. Users click when they trust the source. A brand they have encountered in newsletters, on YouTube, in AI answers, and in community discussions is one they will visit when they are ready to act. Visibility across channels builds the trust that drives the eventual click.
Content that requires human judgment. Reviews, recommendations, strategic analysis, and personal experience cannot be fully replicated by AI. A user searching for “which email tool worked best for a SaaS company with 5,000 contacts” wants a real answer from someone who actually tested the product. That specificity is a traffic driver that AI cannot eliminate.
The businesses that treat AI as a traffic killer will keep cutting content budgets. The businesses that treat it as a distribution shift will adapt their content strategy, expand their channel presence, and find that their total reach holds or grows.
AI is raising the quality threshold. Generic content no longer earns a visit. Specific, credible, experience-based content still does, and it does so consistently.
Audit which of your pages rely on informational queries where AI now dominates. Move investment toward content formats AI cannot replicate: original research, product-specific experience content, and strong editorial opinion. Build presence in channels that bypass search entirely. Track branded search volume and direct traffic as indicators of growing trust, not just session counts from Google.
Traffic is changing shape. Your job is to change with it.