If you’ve noticed fewer people clicking through from Google lately, you’re not imagining it. More of your customers are asking ChatGPT, Google’s AI Overviews, or AI Mode for a recommendation before they ever see a list of blue links, and if your business doesn’t show up in that answer, you don’t exist to them.
The short version: there’s no separate “AI SEO” to learn. Google has said directly that AI Overviews and AI Mode pull from the same index and ranking systems as regular search. If you fix your technical SEO, write answer-first content, and build real credibility signals (reviews, mentions, case studies), you improve your odds in both classic search and AI answers at the same time.
Key takeaways
- ChatGPT now has over 900 million weekly users (as of February 2026), and a growing share of them are asking it questions you’d want your business to answer.
- Google’s own guide says there’s no special schema, no llms.txt file, and no “AI SEO” trick that gets you cited. It’s the same ranking systems.
- Ranking in the top 10-20 for a search term is still the single best predictor of getting cited in an AI answer, but it’s far from guaranteed, even at #1.
- Branded mentions across the web and video mentions correlate more strongly with AI visibility than backlinks do.
- A spam update Google released on September 24, 2026 is rolling out through early October. If your traffic wobbles this week, check this before you panic.
What’s actually different about ranking in AI search right now?
Less than you’d think, and that’s good news if you’re a small business without a big SEO budget. Google’s “Optimizing your website for generative AI features” guide states plainly that AI features run on the same core ranking systems as regular search results. The AI model retrieves candidate pages the way it always has, then writes an answer using pages it’s already decided are relevant, credible, and well-matched to the query.
That means the checklist below isn’t a new discipline bolted onto your existing marketing. It’s the same fundamentals, done well: a site Google can crawl and index without trouble, content that answers a real question directly, and enough outside evidence (reviews, mentions, links) that Google and the AI models trust what your page says. Google explicitly tells site owners to ignore llms.txt files, “chunking” content for AI, and rewriting pages just to please a bot. None of that moves the needle. Depth and trust do.
Here’s the number that matters most: according to Ahrefs’ analysis of roughly 4 million AI Overview citations, a page ranking #1 in regular Google search gets cited in an AI Overview about 53% of the time, while a page sitting at #10 gets cited about 37% of the time. Ranking well still helps a lot, it’s just no longer a guarantee, and that gap has been narrowing as AI models pull more answers from further down the results.
The September 2026 spam update: what it means for you
Google released a spam update on September 24, 2026, its fourth of the year, and said the rollout could take up to two weeks. It’s a global update covering every language, not something aimed at one industry or region.
If your traffic dips or spikes this week or next, this is probably why, and there’s usually nothing to do about it if you’re not running spammy tactics like keyword-stuffed pages, doorway pages, or purchased links. Spam updates target sites gaming the system, not small businesses publishing honest, helpful content. Still, it’s a good week to check Search Console for any manual actions or sudden drops, and to hold off on anything risky (bought links, AI-generated filler content, duplicate location pages) until the dust settles in early October.
The AI search checklist: what to fix this month
Work through these in order. The first few take an afternoon; the later ones are ongoing habits.
1. Check your robots.txt file for blocked AI crawlers. After a site migration or a hosting change, it’s common for AI bots to get blocked by accident. Confirm these are allowed if you want to show up in AI answers:
| Bot | Belongs to |
|---|---|
| GPTBot | OpenAI / ChatGPT |
| ClaudeBot | Anthropic / Claude |
| PerplexityBot | Perplexity |
| Google-Extended | Google’s AI features |
| Applebot-Extended | Apple Intelligence |
| Bingbot | Microsoft Copilot / Bing |
| CCBot | Common Crawl (feeds many AI models) |
We recently walked a client through this exact check after a site migration and found their new host had quietly blocked GPTBot by default in the server config, not the robots.txt file itself. Worth checking both places.
2. Answer the question in the first 40-60 words. Whatever your page title promises, deliver the direct answer immediately, before the background, the story, or the caveats. AI models pull self-contained passages, and if your answer is buried under three paragraphs of throat-clearing, it won’t get pulled.
3. Use question-shaped headings that match how people actually type. “What does a CRM cost?” beats “CRM Pricing Overview” every time, because it mirrors the query.
4. Build your Google Business Profile out completely. Full description, accurate primary category, real photos, and services listed in plain language. This still runs on relevance, distance, and prominence, and you control two of the three.
5. Get reviews on a steady drip, not in bursts. A realistic target is 4 to 8 new reviews a month across Google and one or two industry-specific sites (Houzz or Angi for home services, Healthgrades for healthcare, G2 for software). AI systems and Google both notice unnatural spikes in review volume, so steady beats fast.
6. Add one real case study with numbers. Not a testimonial, a case study: what the client’s situation was, what you did, and what changed. This is the “first-hand experience” signal Google’s own AI guide calls out as the single most valuable kind of content you can publish, and it’s exactly the kind of thing a generic AI-written article can’t fake.
7. Get your business name mentioned, not just linked. A brand mention on a chamber of commerce site, a local news story, or an industry directory correlates more strongly with AI citation than a backlink does. Ahrefs’ research on 75,000 brands puts the correlation for branded web mentions at 0.664 versus just 0.218 for backlinks, and a later update found video mentions (largely YouTube) correlate even higher, around 0.737. You don’t need a link for a mention to count.
8. Keep your cornerstone pages dated and updated. Freshness is a real tie-breaker when an AI model is choosing between two similarly good sources.
9. Don’t chase FAQ schema as a ranking trick. FAQ rich results stopped showing in Google as of May 2026, for everyone, no exceptions. Keep FAQ schema on your page only if it matches text a visitor can actually read, because AI crawlers and Bing still use it. Treat it as a nice-to-have, not a strategy.
10. Check your Core Web Vitals. A slow, janky page is a moderate ranking signal on its own, and a bad experience is a bad experience whether a human or an AI model is the one reading your content.
Common mistakes that keep businesses invisible in AI search
The biggest one is chasing the wrong metric. A lot of business owners assume that if they just rank #1, they’ll automatically show up in every AI answer for that topic. That’s not how it works anymore. Citation and ranking position are correlated, not identical, and that gap has been widening. The businesses getting cited are the ones with the strongest overall trust signals (real reviews, real mentions, real expertise on the page), not just the best on-page keyword optimization.
The second mistake is treating this as a one-time project. A page you built in March and never touched again loses its freshness edge, even if the ranking hasn’t moved. The third is over-investing in schema markup and technical tweaks while skipping the actual credibility-building work (reviews, mentions, case studies) that the evidence says matters more.
How long until you actually see results?
Be honest with yourself here. Long-tail, low-competition terms can move in 8 to 12 weeks if you’re consistent. Broader terms and real authority-building take 6 to 12 months. If you haven’t moved on a target keyword within 90 days, that’s your signal to go back and check whether the page actually answers the question well, not to abandon the whole approach.
Save this checklist
Before you close this tab, pull the quick version:
- Check robots.txt for blocked AI bots
- Answer the question in the first 40-60 words
- Use question-shaped headings
- Fill out your Google Business Profile completely
- Get 4-8 reviews a month, steadily
- Publish one real case study with numbers
- Earn brand mentions, not just links
- Keep cornerstone pages dated and refreshed
- Keep FAQ schema honest, don’t chase it as a hack
- Check your Core Web Vitals
FAQ
Is AI search optimization different from regular SEO?
No. Google has confirmed that AI Overviews and AI Mode use the same core ranking systems as classic search. The work is the same: crawlable pages, answer-first content, and real credibility signals.
Do I need special schema markup to show up in ChatGPT or AI Overviews?
No. Google states plainly that no special schema is required for generative AI search. Keep your existing schema accurate and skip anything sold to you as an “AI schema trick.”
How long does it take to start showing up in AI search results?
Long-tail, specific terms can move in 8 to 12 weeks. Broader terms and real authority-building typically take 6 to 12 months.
Should I block AI bots like GPTBot from crawling my site?
Not if you want to appear in AI answers. Blocking GPTBot, ClaudeBot, PerplexityBot, or similar bots keeps your content out of the tools your customers are increasingly using to find businesses like yours.
By the Easy Design Solutions team. Google Partner, Maryland web services agency.
Want more EDS insights in your Google results? Add us as a preferred source.