A client called us in February with a chart and a problem. Their blog traffic was down 38% year over year. Their conversions were flat. Same posts, same rankings (mostly), same audience. The only thing that had changed was the SERP itself. Google's AI Overviews had moved in above their content like an awning, and the people who used to click through were getting their answer from the awning.
This is now the most common conversation we have with clients who publish anything informational.
We've been watching this pattern since AI Overviews rolled out broadly. The shape is consistent: queries phrased as questions ("how do I," "what is," "why does") lose the most traffic. Queries with commercial intent ("best," "near me," "price") lose less or sometimes gain, because the Overview either refuses to answer or hands users back to the blue links. Branded queries are mostly fine. Local queries are mostly fine. The blood is in the middle: the long tail of educational content that used to bring in a steady trickle of visitors who eventually became customers.
The chart that ruined a quarter
Our client publishes maintenance guides for commercial HVAC equipment. They had a post on detecting a refrigerant leak that ranked #2 for years and brought in roughly 600 visits a month. In Q1 2026, that post averaged 220 visits. The ranking hadn't moved. The Overview had simply absorbed the answer and cited a competitor halfway down.
Here's the thing that kept the client from panicking: their qualified leads from organic search were flat. The visits going away were not the visits that were converting. The contact form fills, the quote requests, the phone calls — those came from a different set of pages, and those pages were holding.
So we made them a deal. Stop measuring blog traffic as the headline number. Start measuring it as a leading indicator. Then change what we publish.
The question Google can't answer in 200 words
This is the test we now run on every editorial idea before it goes on the calendar. If the AI Overview can answer the query in two paragraphs, the post will lose most of its clicks within a year. We can still publish it for topical coverage and EEAT signals, but we should not expect traffic from it.
What survives the test, in our experience:
- Case studies with specific dollar figures, equipment, and outcomes. The Overview can summarize the lesson, but it cannot replace the specifics, and people who care about specifics still click.
- Calculators and interactive tools. An Overview cannot run your math for you.
- Opinion pieces with a clear authorial position. "Here's what we'd actually do" beats "here are five things to consider" every time.
- Comparison content where the comparison requires browsing — galleries, side-by-side specs, long PDFs that need to be downloaded.
- Content that requires the reader to be logged in, geolocated, or otherwise contextualized in a way a public Overview can't replicate.
For the HVAC client we shifted the calendar accordingly. The refrigerant-leak post stays. We're not going to chase a rewrite that outsmarts the Overview; that's not a game we can win. But the next six posts are shaped differently. Three are case studies with named equipment and quoted costs. Two are calculators (one estimates compressor lifespan from runtime data, one converts SEER to operating cost for a given climate zone). One is an opinion piece arguing against a popular maintenance practice. The Overview will struggle with all of them.
What we stopped recommending
Long "ultimate guide" posts. They used to be the workhorse of informational SEO. They are now exactly the format Google's Overview is best at compressing. If you write 4,000 words explaining a concept end to end, you have written training data for the model that will replace you on the SERP. We still write long pieces, but they have to do something the Overview cannot: tell a story, take a position, walk through a specific case, or include something interactive.
The conversion side has not collapsed
This is the part I'll defend hardest: the death of informational SEO has been wildly overstated for businesses that sell things to humans. What's dying is the funnel-top vanity traffic. The bottom of the funnel — the people typing your service plus their city, the people who already know they need a quote — still arrives. We've seen it across every client portfolio we manage. The clients in trouble are the ones whose business model required ad impressions on informational pages. For everyone else, this is a content strategy problem, not a survival problem.
Tell your clients to stop watching the traffic chart and start watching the qualified-lead chart. If both are dropping, you have a real problem. If only the traffic chart is dropping, you have a publishing schedule to rewrite, and you have a few months to do it before the budget conversation gets loud.