The Future of SEO With AI: What the Dated Changes Already Show

Predictions about SEO are cheap. Dated changes from the companies that run search and AI systems are more useful, because you can check them and act on them. Here are the ones that matter most as of October 2026.

None of these say SEO is ending. They say the mechanics moved toward access and away from tricks. If you want the broader argument, read whether SEO is dead. This page sticks to what is documented.

An AI answer about your brand passes through four stages. Knowing which one you are dealing with tells you what you can fix.

1. Parametric. What the model knows with no web access, frozen at a training cutoff. OpenAI documents a 16 February 2026 cutoff for its GPT-5.6 models and 31 August 2025 for the Instant path used in ChatGPT. Nobody changes this stage this quarter.
2. Access. Whether a crawler may fetch and store your page. You and your CDN control this.
3. Retrieval. Whether your page enters the model's context for a given question. The engine controls this, and it rewrites the question before searching.
4. Generation. Which retrieved pages the model uses and cites. The model controls this.

Most published GEO studies measured stage four, on pages already placed in the model's context. A July 2026 critical survey of 45 GEO studies concluded that "already-retrieved content can causally alter its citation or use, but no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability." That is the honest frame for the future of SEO: the proven levers sit at stage two. For a wider look at what changed, see the impact of AI on SEO.

The Cloudflare change is the clearest sign of where things are heading. From 15 September 2026, Cloudflare blocks Training and Agent crawler categories by default on pages that display ads, for the groups listed above, while Search stays allowed by default. Your robots.txt can say yes while your CDN says no, and nothing in your analytics will flag it.

Three controls decide AI access, and each works per bot:

• robots.txt rules are per crawler. OpenAI states that "each setting is independent of the others": GPTBot governs training use, OAI-SearchBot governs ChatGPT search.
• The firewall or CDN is a second gate. OpenAI's help centre says eligibility means allowing OAI-SearchBot and confirming your host or CDN accepts traffic from its published IP ranges. A 403 to that bot is a silent failure.
• Some user-triggered fetchers behave differently. Perplexity documents that Perplexity-User generally ignores robots.txt. The Perplexity visibility factors page covers that split.

A blanket AI block is the common error. This snippet blocks training while keeping ChatGPT search eligibility:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

To see what each crawler actually gets back, the AI Crawler Checker requests a page as each AI bot and shows which ones your firewall turns away.

A lot of AI SEO advice from 2024 and 2025 has lost its footing in primary sources.

• FAQ schema. FAQ rich results were deprecated in May 2026 and the documentation removed in June. Google says structured data isn't required for generative AI search.
• Schema for citations. Ahrefs tracked 1,885 pages adding schema against matched controls and found AI Overview citations fell 4.6%, with no major uplift on any platform. Every page in that study was already heavily cited, so whether schema helps a page enter the answer at all remains untested.
• llms.txt. Google's AI optimization guide says Google Search ignores AI text files. No other engine documents reading one.
• Chunking and AI-friendly writing. The same guide says you don't need to write a specific way for AI search and that there is no ideal page length.

Don't overcorrect. Google says existing markup does no harm, so ripping it out is not a fix either. The accurate line is that markup is not a lever for AI answers. One on-page signal does have some support: stating a specific fact, such as a price or a one-line definition, on a page crawlers can fetch gives an engine a sentence it can quote.

The future of SEO measurement is narrower than most dashboards suggest.

Google

Search Console's generative AI report shows impressions for AI Overviews and AI Mode, broken down by page, country, date and device. It shows no clicks, no CTR and no position. Google has described more metrics only as a future intention. A Google-side bug depressed impressions in both the Search and generative AI reports from around 12 August 2026, so treat trends across that window with care.

The same area holds a control most owners never check. Inclusion in Google's generative features is opt-out, on by default, and child properties inherit their parent's setting. A property can be excluded by inheritance and you would not see it from outside.

Bing

Bing's AI Performance report exposes grounding queries, the phrases the AI used when retrieving cited content. As of August 2026 there was no API for it. Microsoft frames Citation Share as an observational metric, not a ranking.

Rank

Ahrefs found that 37.9% of URLs cited in AI Overviews also appeared in the first 10 results (March 2026). Ahrefs changed its method between studies, so don't read an older, higher figure as a measured decline. Since Google exposes no position data for AI features, any "AI rank" in Google has nothing to be checked against.

Split the work honestly and you stop paying for effort that cannot move anything.

In your control:

• Server-side rendering of the content that matters. Vercel found that ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. Googlebot renders; most AI crawlers don't.
• robots.txt rules per bot, and CDN or firewall settings per bot.
• noindex, nosnippet, max-snippet and data-nosnippet, which keep content out of AI answers if applied, including by accident through a template.
• Index status in Google and Bing, checked separately.
• Consistent facts across your own pages. If pricing reads differently on three pages, an engine has three answers to pick from. See what to do when ChatGPT gets your pricing wrong.

Not in your control:

• What the model memorised before its cutoff.
• How the engine rewrites the question and which pages it retrieves.
• Where your page sits in the model's context, and which sources it cites.
• Brand prominence. Large brands dominate AI citations, and no page edit changes that this quarter. Being named on sites the engines already cite is associated with appearing in answers, with no promised effect.

Order of work

1. Test access per bot. Read robots.txt for OAI-SearchBot, PerplexityBot, Claude-SearchBot, Meta-WebIndexer and Googlebot. Then request a page as each one and confirm a 200, not a 403 or a challenge page.
2. Check your CDN defaults. If you use Cloudflare, review which AI crawler categories it now blocks.
3. Diff raw HTML against the rendered page. Anything that only appears after JavaScript runs, such as prices or plan names, needs to move into the server response.
4. Search for accidental suppression. Look for noindex, nosnippet and data-nosnippet in templates, and check the generative AI setting in Search Console.
5. Confirm index status in both Google and Bing.
6. Fix your facts. Make each price, product name and key fact identical everywhere you state it. After a rename or relaunch, the rebrand checklist for AI answers sets out the order.

Common mistakes

• Allowing GPTBot to "get into ChatGPT". GPTBot governs training only.
• Treating one index as all of them. In Otterly's 2026 experiment, Google deindexed two AI-generated test sites while Bing kept them, and Microsoft Copilot kept citing them.
• Mass-publishing thin pages. Google's spam policy names generating many pages with AI tools without adding value as scaled content abuse.
• Reporting one blended AI score. Engines differ too much for an average to tell you what to fix.

The useful habit is watching dated changes from the operators, not commentary about them. Keep a short list and check it monthly.

• Your CDN's bot settings. Defaults can move again, and they apply to your zone whether you read the announcement or not.
• Vendor crawler pages. OpenAI, Anthropic, Perplexity, Google and Meta each document their bots. New user agents and changed purposes appear there first.
• Search Console. Watch for new metrics in the generative AI report and for the inclusion control reaching your property.
• Bing Webmaster Tools. Export grounding queries now, and watch for the API Microsoft has indicated is coming.
• Model cutoffs. Each new model moves the parametric boundary. Content published after a cutoff is absent from that model's answers when it doesn't search.
• What each engine says about you. Check engine by engine, for example with a free check of what Copilot says about you. The free checker on this site gives a snapshot for up to three buyer questions, which is a fair starting baseline.

For the parts of classic SEO that still carry weight, read SEO in the age of AI.