Is AI search killing my SEO traffic?

Q: How do I know if AI is causing my traffic drop?

Isolate the signature. Segment Search Console into informational, transactional, and branded queries; AI erosion hits the informational bucket first. If impressions are flat or rising while clicks fall, the query is being answered without a visit. If impressions also fell, you lost ranking, not clicks to AI. Then search the term and confirm an AI Overview actually appears above the organic results.

Q: Which queries lose the most clicks to AI Overviews?

Definition-style and 'what is a good X' queries are the clearest losers. The visitor wanted a number or a quick fact, got it inline, and never clicked. These were low-intent, low-converting clicks to begin with. Commercial terms like pricing, comparison, and 'near me' queries rarely trigger an Overview and keep sending clicks.

Q: Does being cited by AI actually send traffic?

Yes, though less than a top organic ranking used to. AI Overviews and chat answers in ChatGPT, Perplexity, Gemini, Claude, and DeepSeek link out to their sources, and named citations still pull clicks and appear in front of buyers researching inside those engines. The value is shifting from raw click volume toward being the source the model trusts.

Question

Is AI search killing my SEO traffic?

Maciej Grabek · Accepted Answer

Sometimes, but rarely the way headlines claim. AI Overviews and chat answers absorb some informational, zero-click queries, so your impressions can hold steady while clicks fall on top-of-funnel terms. The effect is uneven: transactional and branded searches barely move, and being the source AI cites still sends traffic. Before you blame AI, segment your Search Console data by query type and check which terms actually show an AI Overview. What is actually happening to your clicks Two takes dominate the conversation and both are wrong. "SEO is dead" ignores that most commercial, transactional, and branded searches still send clicks the same way they did in 2022. "Nothing changed" ignores that a real slice of informational traffic now resolves without a visit. The truth is uneven and query-specific. Here is the mechanism. On many informational queries, Google now shows an AI Overview above the organic results and answers the question in place. Your page can still appear in or below the Overview, so Search Console logs an impression, but the user has no reason to click. That produces a distinct signature: impressions flat or rising, clicks falling, click-through rate dropping, concentrated on top-of-funnel informational terms. The same pull happens in chat, where someone asks ChatGPT or Perplexity directly and never opens a search results page at all. How to tell if AI is actually the cause Traffic dips have many causes: a core update, seasonality, a tracking bug, a competitor outranking you. Blaming AI without checking is how you end up rewriting a content strategy to fix a broken analytics tag. Isolate the signature first. Segment Search Console by query type. Split informational queries ("how", "what", "why", "best way to") from transactional ("buy", "pricing", "near me", "vs") and branded (your own name). AI erosion shows up in the informational bucket first and barely touches the other two. Compare clicks against impressions. If impressions are flat or rising while clicks fall, the query is being answered without a visit - the AI Overview signature. If impressions also dropped, you lost ranking, not clicks to AI. Confirm the Overview exists. Search each suspect term and look for the AI Overview panel above the organic list. No Overview means AI is not the culprit for that keyword. Check the timeline. AI Overviews rolled out broadly across US queries in mid-2024. A dip that predates that window is not an AI story. One query type that is losing clicks to AI Definition-style and "what is a good X" queries are the clearest losers. They ask for a fact or a number, which is exactly what an AI Overview or a chat answer is built to hand over inline. Ask Perplexity "what is a good email open rate for B2B" and it answers in place: a range, a couple of caveats, and three cited sources, with no reason to click through. Google's AI Overview does the same thing above the organic results for the identical query. Notice what these queries have in common: they were never high-intent. The visitor wanted a number, got it, and left - a click that rarely converted anyway. Losing it stings in the dashboard but costs almost nothing in pipeline. The queries that still send valuable clicks - "email marketing platform pricing", "Mailchimp vs Klaviyo", "book a demo" - are the ones AI mostly leaves alone, because a model summary cannot close a purchase decision. What to do instead of panicking Reprice the loss. If the eroding terms are top-of-funnel and low-converting, a click drop there is not a revenue drop. Check assisted conversions and pipeline before you overhaul anything. Compete to be the cited source. AI Overviews and chat answers link out, and being named still sends traffic and puts you in front of buyers researching inside ChatGPT, Perplexity, Gemini, Claude, and DeepSeek. That is what GEO (generative engine optimization) targets. Shift content weight toward buyer-stage pages AI cannot fully replace: comparisons, pricing, original data, and specific product detail. Those hold their clicks and feed the citations. Measure your AI visibility, not just your rankings. Run your key prompts in each of the five engines once a month and log whether you are named and cited. Rankings alone no longer tell you whether buyers are finding you. Running every prompt across five engines by hand gets old fast. avisibli automates the tracking - which engines cite you, for which prompts, and how that shifts week to week - so you can watch AI visibility next to your Search Console clicks instead of guessing which one is moving. avisibli is the GEO platform that publishes this answer library. Self-references are limited to topics where a tool-based answer is genuinely useful to readers.