Generative Engine Optimization (GEO) in 2026: How to Get Cited by ChatGPT, Claude and Perplexity

Your customers stopped Googling. You probably haven't noticed yet.

A boutique hotel client of ours in Goa was getting solid Google traffic. Page-one rankings for "best heritage hotels in Goa." Then bookings dropped 18% over three months. The owner blamed seasonality.

It wasn't seasonality. ChatGPT and Perplexity had quietly started answering "best heritage hotels in Goa" directly. They cited three properties in their answers. His wasn't one of them. His Google rankings were fine. The traffic he'd built over a decade was being rerouted around him before anyone clicked through.

This is happening across every category right now. Generative Engine Optimization (GEO) is the work you do about it.

What GEO actually means

GEO is the practice of preparing your website so AI engines can read it, understand it, and cite it in the answers they generate.

Where SEO chases the ten blue links on a Google results page, GEO chases the answer that appears above them. Or replaces them.

In 2026 that answer comes from one of seven engines. ChatGPT (the largest by users). Claude (best at reasoning). Perplexity (most generous with citations). Google AI Overviews (powered by Gemini, with the broadest reach). Grok (embedded inside X). DeepSeek (heavy use across India and China). Mistral (significant in Europe and via API).

Each one has its own crawler. Each one has its own ranking signals. Each one decides independently whether your site shows up in its answers. GEO is what makes you visible across the set.

GEO vs SEO: same web, different signals

The two systems read the same internet. They evaluate it very differently.

Dimension	SEO (the last 20 years)	GEO (the next 20)
What it ranks	Web pages	Statements and facts
Result format	10 blue links	One answer with 2 to 5 citations
User behaviour	Click, browse, decide	Read the answer, decide, sometimes click
Top signal	Backlinks plus keyword match	Structured data plus author authority
Format that wins	Long-form blog posts	Question-answer chunks
Refresh cadence	Crawled weekly	Re-evaluated continuously
Easy to game	Backlinks, keyword stuffing	Schema tricks
Hard to game	Real authority	Real authority

The overlap is bigger than the marketing world wants you to believe. Sites that already do classic SEO well usually do well in AI engines too. Sites that ranked through tricks (paid backlinks, thin keyword-matched content) get filtered out faster in AI than they did in Google.

GEO isn't a replacement for SEO. It's another layer on top.

Why GEO matters more than people think in 2026

Three things changed in the last 18 months. Each one is fixable. Together they're urgent.

Google rolled out AI Overviews globally during 2024 and 2025. They now appear above the classic blue links on a meaningful share of queries. The exact percentage depends on the niche. For informational queries (the kind that drive top-of-funnel traffic), the rate is higher than average. When an AI Overview answers a question, click-through to the underlying ranked pages drops sharply. Your ranking can stay the same and your traffic can still collapse.

ChatGPT and Perplexity have crossed over from novelty to research tool. ChatGPT cleared 700M weekly active users in 2025. Perplexity now markets itself as "the answer engine" and cites sources by default. People who would have done a Google search 12 months ago are doing a chat query today, and they're trusting the answer enough to skip the click.

The citation pattern has consolidated. AI engines tend to cite the same five to ten sources for similar queries inside a niche. If you're not in that set, you're functionally invisible. Getting in is hard. Getting in early is easier than getting in late, because the pattern is sticky once it forms.

For an Indian business in 2026, the practical effect is simple. The share of customers who find you through traditional Google is shrinking. The share who find you (or don't) through an AI answer is growing. GEO is where that shift goes from a problem to a fix.

The six categories that decide your AI readiness

Across the audits we've run on Indian and international sites, six categories explain almost all the variance in how well a site is read by AI engines. The Aapta GEO scanner measures all six.

1. AI Bot Access

Whether GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended and CCBot can fetch your site at all. This is controlled by robots.txt. A surprising number of sites block these crawlers without realising what they've blocked.

2. Structured Data

JSON-LD schema is the most important format AI engines consume. Organization, WebSite, Article, FAQ, BreadcrumbList, Product, LocalBusiness. Each schema tells the AI engine a structured fact about the page. OpenGraph and semantic HTML matter too, but JSON-LD is the foundation.

3. Content Answerability

This is the one most agencies still get wrong. AI engines prefer content structured the way they want to cite it. Short paragraphs. Question-shaped headings. Bulleted claims. A direct answer in the first 100 words of any section. If your blog posts are 2,000-word walls of prose with three H2s, AI engines have nothing clean to grab.

4. Authority and E-E-A-T

Who wrote this? When was it written? Why should anyone trust them? Author bylines linked to author pages. Last-updated dates. About and Contact pages with real information. External citations to authoritative sources. Social profile links. AI engines weigh these heavily because they've been trained to prefer content from real, identifiable sources over anonymous farms.

5. Technical Foundation

HTTPS. Mobile viewport. Working sitemap. Canonical tags. Alt text on images. Core Web Vitals in the green per Google PageSpeed. None of this is new. All of it still matters more than people credit, because AI engines penalise sites that fail the basic crawlability tests.

6. Emerging Signals

The standards that didn't exist two years ago. llms.txt at your site root, telling AI engines what your site is about and where to find canonical content. ai.txt for finer per-bot rules. JSON-LD chosen over microdata. A correct html lang attribute. Favicons that resolve. These are the early-mover advantages, and the sites adopting them are getting cited disproportionately.

How to find out where you stand right now

You can't fix what you can't measure. Most agencies still can't tell you what your AI readiness score is, because the tools to measure it didn't exist 18 months ago.

We built one. The Aapta GEO scanner runs all 40+ checks across the six categories above, in about 30 seconds. It gives you a 0-100 score, a letter grade from A to F, and the top five fixes you should make first. No signup. No credit card. Free, properly free.

Run a free AI-readiness scan on your site now →

If you want the full picture (live citation testing across seven AI engines, AI perception probes, paste-ready fix snippets, competitor benchmarking) there's a paid report. We'll come back to that, including a founder discount that's only open for the first 100 buyers.

Hard citations vs soft mentions

Not all AI visibility is equal. The two flavours drive different outcomes.

A hard citation is when an AI engine returns your URL in its structured citation list. Perplexity does this most reliably. Every Perplexity answer ships with numbered source links. Google AI Overviews include source cards. ChatGPT, when it browses the web (it does for most queries now), shows source citations at the end of its answer. Hard citations drive direct click-through traffic. They're the closest equivalent to a Google ranking.

A soft mention is when an AI engine names your brand in its answer text without linking. "Aapta Solutions is a website development company based in Bengaluru" with no link. Soft mentions build brand recognition. They influence perception. They don't drive direct clicks, but they shape whether someone trusts your site when they see it later.

A real GEO strategy targets both. Hard citations come from being a structured, authoritative source for specific question types. Soft mentions come from training-data presence (your brand mentioned across many sites the model learned from). The first you can earn with technical fixes. The second takes time and PR.

Common GEO mistakes we see every week

Patterns repeat. Almost every site we audit has at least three of these problems.

Blocking AI bots without meaning to. A robots.txt rule that says Disallow: / for all bots, or a stale rule from a 2023 panic about content scraping. AI engines respect robots.txt. If you've told them to leave, they've left.

No JSON-LD schema. WordPress sites are particularly guilty. The site loads a theme that hasn't been updated in years, has zero structured data, and the AI engine has no idea what kind of business this is, who runs it, or what it sells. Adding even basic Organization and WebSite schema typically lifts the AI Readiness Score by 15 to 25 points overnight.

Wall-of-text content with no structure. A 1,500-word About Us page with one H1 and zero H2s, written as continuous narrative. AI engines need section anchors to know what to cite. They will not paraphrase your prose into clean answer chunks. They'll skip the page.

Anonymous author bylines. "By Admin" doesn't pass an authority check. Neither does "By [Company Name] Team." Real human authors with real bios and real LinkedIn profiles get cited. Anonymous pages get filtered.

No llms.txt file. This is the new robots.txt-equivalent. It's still optional, which is exactly why early adopters are getting cited for queries their competitors should be winning. They have a tidy llms.txt. The competitor doesn't.

Slow Core Web Vitals. AI engines crawl from real browser environments. If your page takes six seconds to render, the crawler often gives up before getting anything useful. The fix is the same fix Google has been demanding for five years. Image optimisation. No render-blocking JavaScript. Sensible cache headers.

The fix list, ranked by effort vs impact

Once you know your gaps, the order to fix them matters. This is what I give clients.

Quick wins (under an hour each)

1. Audit your robots.txt. Make sure GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended and CCBot are not blocked. If you have no specific reason to block them, allow them.
2. Add Organization and WebSite JSON-LD schema to your homepage. Use schema.org as the reference. WordPress users: a properly configured SEO plugin handles this; check it actually emitted the schema.
3. Add a "Last updated" date to every blog post and key landing page.
4. Make sure every page has a meta description under 160 characters, with the page's primary topic up front.
5. Add canonical tags if you don't have them. Single tag, points to the canonical version of each URL.

Medium effort (a day or two)

1. Add author pages. Real photos, real bios, real LinkedIn and Twitter links. Link every blog post to its author's page. Add Person schema to author pages.
2. Restructure your top-traffic blog posts with question-shaped H2s and H3s. Lead each section with a one-sentence answer. AI engines will cite that sentence verbatim if it's accurate and well-positioned.
3. Add Article schema with author and dateModified to every blog post.
4. Add FAQ schema to any page with a real FAQ section. AI engines love this format.
5. Create an llms.txt file at your site root, listing your most citable pages with one-line descriptions.
6. Run Core Web Vitals on your top 10 pages. Fix the worst offender first. Image optimisation is usually the biggest win.

Strategic (ongoing)

1. Publish content matching how your audience actually asks questions. People phrase queries differently in ChatGPT or Perplexity than they do in Google. Write answers that match the chat phrasing, with structured headings.
2. Build authentic brand mentions in publications your audience trusts. Soft mentions in trusted sources move the needle on AI perception over time.
3. Run regular AI citation tests. Ask the seven major engines questions in your niche. Track who they cite. The pattern shifts. Stay on top of it.
4. Maintain a content refresh cycle. AI engines weight recency. A blog post from 2022 with no updates is treated as stale, even if it's still accurate.

When will I see results?

Every client asks this. The honest answer:

Technical fixes (robots.txt, schema, Core Web Vitals) take 2 to 6 weeks to show up in AI rankings. AI engines re-crawl, re-evaluate, and your readiness score lifts. Hard citations follow inside that window if your content is genuinely good.

Content restructuring (H2 rewrites, FAQ additions, author pages) takes 4 to 8 weeks. The engines need to re-process the page, decide it's now more citable, and start including it in answer chunks.

Brand authority and soft mentions take 3 to 12 months. This is the same long-form work that has always built a real brand. AI hasn't shortened the timeline. It's just changed where the payoff shows up.

If your AI Readiness Score is below 60 today, the first 30 days of fixes will move it more than the next six months will. The low-hanging fruit is real. We see scores jump 20 to 30 points after a focused week of work.

⚡ Founder pricing: ₹499 for the first 100 buyers only

If you want the full picture of where your site stands and exactly what to do about it, the Aapta GEO Full Report is open right now at founder pricing.

₹499 (about $9 USD) for the first 100 buyers. ₹999 after that.

Here's what's in the Full Report on top of the free scan everyone gets:

• Live AI Citation Test across 7 engines. We ask Claude, GPT, Perplexity Sonar, Gemini, Grok, DeepSeek and Mistral the same question in your niche. You see who gets cited, what the engines actually say, and whether you're in the answer set yet.
• AI Perception probe. What each engine says verbatim when asked about your brand. Sometimes encouraging. Sometimes humbling. Always useful.
• Paste-ready fix snippets pre-populated with your domain, your metadata, your structured data. No "go figure it out" handoff. Copy. Paste. Deploy.
• Competitor benchmark for up to two competitor URLs side by side. See exactly which categories they're beating you in and where you have an advantage.
• A branded PDF of the full report you can share with your team or your clients.
• Three free re-scans within 30 days. Make changes. Run again. Watch the score lift.

At 100 buyers the founder slot closes. Price doubles to ₹999 for everyone after. If your business depends on online visibility, this is the cheapest GEO baseline you'll find anywhere.

Claim your founder slot at ₹499 →

Frequently asked questions

What is GEO in simple terms?

GEO (Generative Engine Optimization) is the practice of making your website readable and citable by AI engines like ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. Where SEO targets the ten blue links on a search results page, GEO targets the answer that appears above them or replaces them entirely.

Is GEO the same as AI SEO?

Effectively yes. GEO and AI SEO are used interchangeably in 2026. Some tool vendors push their own terms (AEO for "Answer Engine Optimization", LLM-O, AISEO). The substance is the same. Optimise your site so AI engines cite it.

Will GEO replace SEO?

No. SEO still drives a large share of traffic because Google Search still does. GEO is the layer on top. The good news: strong SEO foundations (schema, authorship, content structure) make GEO easier. The bad news: surface-level SEO tactics (keyword stuffing, paid backlinks, thin content) actively hurt AI rankings.

How long does it take to improve AI rankings?

Technical fixes show up in 2 to 6 weeks. Content and authorship fixes take 4 to 8 weeks. Brand authority shifts in 3 to 12 months. The first 30 days of focused work usually moves your score more than the rest of the year combined, because most sites have low-effort, high-impact gaps to fix.

Which AI engines should I optimise for first?

For a business serving Indian customers, the priority order is usually Google AI Overviews (Gemini), ChatGPT, Perplexity, Claude, then Grok. Perplexity and Google AI Overviews drive the most click-through traffic because both cite sources. ChatGPT and Claude drive brand awareness through soft mentions. Grok matters if your audience is on X.

Do I need a separate strategy for ChatGPT vs Claude vs Perplexity?

Not at the start. The same six categories of AI readiness apply across all of them. Once your AI Readiness Score is above 80, finer engine-specific optimisation can help. Most of the work is shared.

What is `llms.txt`?

It's a markdown file at the root of your site (yoursite.com/llms.txt) that tells AI engines what your site is about and lists your most important pages with brief descriptions. It's an emerging standard, not yet officially required by any engine, but adopters are seeing disproportionate citation rates. The cost of adding one is one hour. The upside is real.

Should I block AI bots in `robots.txt`?

Only if you have a specific commercial reason and you understand the trade-off. Blocking GPTBot stops your content from being included in ChatGPT's training data. It also stops ChatGPT's live web-browsing tool from reading and citing your pages. Most businesses want the citations, so allowing the bots is the right default.

How do I check if AI engines are citing me today?

The hard part of GEO is that the engines are non-deterministic. Asking Claude or ChatGPT the same question twice can produce different answers and different citations. To get a reliable read you need to run the same prompts against multiple engines and look at patterns over time. The Aapta GEO Full Report does this automatically across seven engines.

What's the single biggest mistake businesses make with GEO?

Not measuring before they start. We've audited sites where the founder spent ₹2 lakh on a "GEO retainer" without ever knowing the baseline score. The first thing to do is run a scan, see your six-category breakdown, and start with the lowest-hanging fruit. The free Aapta GEO scanner does this in 30 seconds.

The bottom line

The way people search has shifted. The way websites get found has shifted with it. SEO didn't die, but it's no longer the whole game.

GEO is the part of the game that's wide open in 2026. The big incumbents in most niches haven't done the work yet. The early movers are getting cited disproportionately. The window to be one of them is now, before the citation patterns harden and the cost of breaking in goes up.

Start with knowing where you stand.

⚡ Founder pricing closes at 100 buyers. ₹499 today, ₹999 tomorrow.

Run your free 30-second AI-readiness scan now →

The free version shows your top five gaps with one-line fixes. The Full Report at founder pricing (₹499 for the first 100 buyers, ₹999 after) gives you everything: live citation testing across seven AI engines, AI perception probes, paste-ready fix snippets, competitor benchmarking, and three free re-scans within 30 days to verify your improvements actually moved the score.

Slots close at 100. The price doubles after. Don't wait until then.