More than 50 million people now use ChatGPT weekly for search queries. Perplexity reached 100 million users. Google itself has started showing AI-generated summaries at the top of results — before the first organic link.
Search behavior is shifting. And the optimization strategies most businesses are investing in aren't keeping up.
Here's the uncomfortable truth: your SEO ranking on Google has almost no correlation with whether you get cited by AI systems. They operate on entirely different logic. A business that's invisible on page 3 of Google could be regularly cited by ChatGPT — and a business ranking #1 might never appear in an AI answer.
This piece explains why, and what you can do about it.
The Gap Nobody Is Talking About
When someone searches Google for "best Slack bot developer London," they see ten blue links. They click one or two. The business that ranks #1 has a massive advantage — they get roughly 30% of all clicks on that query.
When someone asks ChatGPT the same question, they get a synthesized answer. One paragraph. Two or three names cited, with brief descriptions of each. The winner of that interaction isn't ranked #1 in an algorithm — they're cited by a language model that made a judgment call based on what it found when it crawled your site.
These are fundamentally different systems. The ranking signals for Google — backlinks, domain authority, keyword density, site speed — are largely irrelevant to whether an AI cites you. AI systems care about different things: clarity, structure, specificity, and citability.
Traditional SEO vs. GEO: What Each Engine Actually Cares About
| What Matters | Traditional SEO (Google) | GEO (AI Search) |
|---|---|---|
| Primary signal | Backlinks & domain authority | Content clarity & extractability |
| Content format | Long-form, keyword-rich | Specific, citable, structured |
| Structured data | Nice to have (rich snippets) | Essential (primary data source) |
| Freshness | Matters for news, less for services | Actively deprioritizes undated content |
| Bot access | Googlebot only | 7+ AI crawlers, each with own rules |
| Site index | sitemap.xml | llms.txt (new standard) |
| Social proof | Stars in search results (schema) | Requires Review schema to be cited |
| Optimization timeline | 3–6 months to see movement | 2–4 weeks after re-indexing |
The two disciplines overlap in some areas — both benefit from well-structured content and proper technical setup. But the specific mechanics are different enough that you can't assume your SEO work transfers to AI visibility.
What AI Models Actually Use to Decide Who to Cite
Based on the Princeton KDD 2024 research on generative engine optimization and practical testing, here are the signals that matter most:
| Signal | What It Means | Impact |
|---|---|---|
| Specific statistics | Actual numbers, percentages, prices | +115% citation rate |
| Authoritative citations | Links to research, industry sources | +80% citation rate |
| FAQ/Q&A structure | Question-answer format AI can extract | High — mirrors user query format |
| Schema markup | JSON-LD structured data | High — primary extraction source |
| llms.txt | Machine-readable site summary | Medium — improves entity understanding |
| Recency signals | <time> elements, update dates |
Medium — prevents deprioritization |
| Entity clarity | Clear name, location, service category | Medium — enables correct categorization |
The citation rate numbers come from the Princeton study (Aggarwal et al., 2024) which tested specific GEO techniques across multiple generative engines. The +115% and +80% figures refer to improvements in the probability of being cited versus a baseline page without those optimizations.
The 5-Minute Self-Audit
Before investing any time in GEO work, run this test. It takes 5 minutes and tells you exactly where you stand.
- Open Perplexity.ai. Search: "[your service type] in [your city]". Do you appear? Note which competitors do.
- Open ChatGPT. Ask: "I need [your service] for my [type of business]. Can you recommend someone?". Same question.
- Check your robots.txt. Visit yourdomain.com/robots.txt. Does it mention GPTBot, ClaudeBot, PerplexityBot? If not, AI crawlers may not be indexing you.
- Check for llms.txt. Visit yourdomain.com/llms.txt. If you get a 404, that's one of the first things to fix.
- Check for schema markup. View source on your homepage and search for
application/ld+json. No results? You have no structured data.
If you failed steps 3, 4, and 5, that's good news — they're the easiest fixes and they're very likely to improve your citation rate.
The New Playbook
GEO isn't replacing SEO. You still need Google visibility. But for B2B service businesses especially — where a single new client from an AI recommendation could be worth $5K–15K — the return on GEO investment is disproportionately high.
The new playbook for 2026 looks like this:
- Keep doing SEO. Google still drives most web traffic. Don't abandon it.
- Add GEO as a parallel track. It requires a one-time structural investment (robots.txt, llms.txt, schema markup) and then ongoing content discipline.
- Write for extractability. Every page should have at least one paragraph that could be directly quoted in an AI answer: a definition, a fact, a specific claim.
- Update and date your content. Quarterly updates with visible dates signal freshness to AI systems.
- Measure citation rate, not just rankings. Set up a monthly check: run 5–10 test queries in Perplexity and ChatGPT and track whether you're appearing.
The businesses that are investing in this now are building an advantage that will be much harder to close in two years, when everyone knows about GEO and the structural gap no longer exists.
One More Thing
I want to be direct about something: GEO is not magic. It's not going to turn a bad service into a cited one. AI systems are surprisingly good at evaluating the quality and specificity of content. If your website has vague copy and no real substance, structured data won't save you.
What GEO does is remove the structural barriers that prevent a good business from being seen. If your work is solid and your clients are happy, GEO helps the AI systems understand that — and pass it on to people asking the right questions.
That's worth something.
