AI Overviews Ate Your Blog Traffic. Now What?

AI Overviews Ate Your Blog Traffic. Now What?

Website Dreamwork

Key takeaways

  • AI Overviews now appear on roughly 60% of searches, and on those queries, clicks to websites dropped about 58%.
  • Informational queries got crushed hardest. “How long does an oil change take” no longer needs a website visit.
  • Local “near me” and transactional queries are largely safe. AI can’t fix your dishwasher.
  • The pivot isn’t “stop writing blog posts.” It’s “write fewer, deeper ones that are designed to be cited.”

If your blog traffic started slipping sometime in late 2025 and never recovered, you’re not imagining it. Google’s AI Overviews — the AI-generated answer box at the top of search results — quietly rewired a huge chunk of the search-to-website pipeline. The traffic didn’t go to a competitor. It went to the box.

This is the part where most marketers panic. It’s also the part where the strategy that worked is the one you stop doing. Let’s actually look at what changed and what works now.

What the data is actually showing

Across the small-business sites I work on in East Texas, the pattern is roughly the same:

  • Posts written to answer simple questions (“how often should I service my AC?”) — traffic down 50% to 80%.
  • Posts written to compare options (“Wix vs WordPress for a contractor”) — traffic down maybe 20%, but the visitors that do come are more serious.
  • Local + service pages (“HVAC repair in Longview”) — barely moved. Some are up.
  • Booking, contact, and pricing pages — flat or up. AI doesn’t book your appointment for you.

The takeaway: AI Overviews ate the casual, low-intent informational traffic. They haven’t touched (much) the traffic that was actually converting into customers.

The traffic you lost was never going to call you. The traffic that calls you mostly didn’t change.

Why this is actually… not the end of the world

If you measure your blog by pageviews, this looks catastrophic. If you measure it by leads, it’s much less dramatic. The vast majority of blog visitors on a small-business site were never going to become customers — they were people looking up a definition or a quick fact. AI Overviews are honestly better at that than your blog post was.

What didn’t change: when someone is ready to do something — book an appointment, request a quote, drive to a location — they still click. They click your contact page, your service page, your map listing.

The new shape of small-business content

Here’s the strategy that’s working in 2026.

1. Fewer posts, deeper posts

The old playbook said “publish two posts a week to feed Google.” That playbook is dead. The new playbook says “publish one really good post a month that’s the best thing on the internet for that query, and that AI engines will cite.”

A post that AI cites in an Overview gets shown to thousands of searchers. Some click the citation. Some don’t. But the brand recognition compounds.

2. Structure for citation, not just reading

AI engines like clean structure. Question-shaped headings (H2/H3 that match how a real person would ask the question). Short, direct answers near the top of each section. Bulleted lists. Tables. The kind of clean structure a tired human also likes.

Practical: if your subheadings are clever and abstract (“The Hidden Truth About HVAC”), rewrite them as questions (“How often does an HVAC system actually need service?”). AI engines reward the literal phrasing.

3. First-person, place-specific, opinionated

AI engines are still very bad at first-hand experience. They can summarize ten generic blog posts about local SEO. They can’t write what it’s like to fix a small-business website in Longview after a storm took out the SWEPCO grid. That kind of content — specific, local, opinionated — is what AI engines cite when they need to be more than a generic summary.

4. Stop writing the easy ones

“What is a backlink?” doesn’t need another blog post. AI Overviews own that query and they always will. Spend your effort on the questions that don’t have a clean, factual answer — the trade-offs, the local angles, the opinion pieces, the case studies.

Old playbook (2018–2023) New playbook (2026)
2 posts/week, 800 words each 2 posts/month, 1,500+ words each
Target high-volume keywords Target high-intent keywords
Clever, abstract headlines Literal, question-shaped headings
Generic “ultimate guide” Specific, local, opinion-led
Measure by pageviews Measure by leads and citations

The local angle protects you

Here’s the thing that doesn’t get said enough in the doom-loop articles: local content has been one of the most AI-resistant categories. Nobody asks AI “fix my plumbing.” They ask Google “plumber near me” — and the map pack and local website results still get the call.

If your blog is writing for a local audience about local-specific things, you’re way less exposed than a national lifestyle blog is. A post called “What Longview homeowners should ask before hiring a roofer after a hail storm” is not a post AI is going to swallow whole. It can summarize what to ask. It can’t be the Longview-specific authority that ranks for that phrase.

What to actually do this month

  1. Pull your top 20 blog posts from Search Console by last 12 months of clicks.
  2. Compare the last 90 days to the prior 90 days. Which posts collapsed? Which held steady?
  3. The ones that collapsed are usually informational posts AI now answers. You can either let them go, consolidate them into a deeper guide, or rewrite them to add the human-specific value AI can’t.
  4. The ones that held steady are the template for what to write more of. Study what they have in common.
  5. Add FAQ schema to your service pages and the deeper posts. AI engines parse and cite FAQ blocks heavily.
  6. Stop scheduling posts you don’t believe in. A single great post a month beats four mediocre ones.
Don’t: generate replacement content with AI. AI engines actively de-prioritize AI-generated content when they’re citing sources. Generated content is also what’s flooding the very pool you’re trying to stand out in.

Where people go wrong (and when to call a pro)

The common reaction to an AI-driven traffic drop is to either give up on content entirely or to double the publishing rate. Both are wrong. The correct response is the harder one: write less, but make each piece genuinely valuable, specific, and citable. If you can’t tell which of your pages are still earning their keep, a Search Console audit by someone who’s looked at fifty small-business sites is usually more useful than another month of guessing. The good news is the businesses that respond well to this shift end up with stronger, less-noisy sites than they had before.

Frequently Asked Questions

Should I delete my old blog posts?

Not delete — consolidate. Take five thin posts on related topics and merge them into one deep guide with the older URLs redirecting in. That’s almost always a better signal to search engines than mass deletion.

Is SEO dead?

No. SEO is now SEO + GEO (generative engine optimization), and a lot of the underlying work is the same — be useful, be specific, be structured. The metrics changed. The fundamentals didn’t.

Do AI Overviews actually drive any clicks?

Some, yes. Being cited in an Overview drives roughly the same clicks as a position 5–8 organic result — not nothing, but not what a #1 ranking used to deliver. The bigger benefit is brand exposure.

How do I show up in an AI Overview?

There’s no checkbox. But these patterns help: clean H2/H3 structure with literal phrasing, FAQ schema, first-hand experience that can’t be summarized from elsewhere, and being a recognizable domain that other sites already cite.

If your traffic has been sliding and you’re not sure what’s worth saving, we can audit your top pages and map out a content plan that fits the new shape of search.

Get a Free Quote

Leave a Reply

Your email address will not be published. Required fields are marked *