Baseline
Ask each assistant the questions your buyers would ask. Record the answers and the citations. Almost always more sobering than expected.
AI & Next-Gen Search · Delhi NCR & remote
When someone asks ChatGPT for a recommendation.
A model names three or four businesses. It does not run an auction and it does not check who ranks first. It draws on what it has read, and on what it can verify from more than one place.
Short answer
GEO, or generative engine optimisation, is the work of becoming a source that generative assistants such as ChatGPT, Gemini, Perplexity and Copilot draw on and cite. It depends less on ranking and more on being clearly described, consistently identified across the web, and corroborated by sources other than your own site.
ChatGPT, Gemini, Perplexity
Third-party corroboration
No control, only influence
Monitored monthly
Free GEO audit
Get a free quote
Four details. A real reply the same day, from me.
Do the assistants know you exist?
Get a Free Quote
WhatsApp
01 The premise
A different kind of visibility.
Ask an assistant to recommend a digital marketing consultant in Delhi and it will name a handful. That list was not produced by a ranking algorithm you can study. It came from what the model absorbed during training, plus whatever it retrieved from the live web at the moment you asked. Neither is something you can buy into.
What you can influence is whether the model has clear, consistent, corroborated information about you. Models are conservative about naming specific businesses, and for good reason — recommending a business that does not exist is a serious failure. The practical consequence is that they favour entities that are described the same way in several independent places.
This makes GEO look less like classic SEO and more like public relations combined with data hygiene. Your own site matters, but it matters partly as the thing other sources agree with. A business described identically on its site, its Google profile, LinkedIn, a directory and a news mention is far easier for a model to name with confidence than one that appears three different ways.
I will be direct about the limits: nobody outside these companies can measure this precisely, and the field is young enough that some of what is confidently sold as GEO is guesswork. The underlying work — being clearly described, genuinely credible and consistently identified — is worth doing regardless, which is the honest reason to do it.
02 The work
What the work actually is.
Roughly in order of how much difference each makes.
The work
- Entity consistency. Name, description, location and services identical across your site, Google Business Profile, LinkedIn, directories and every profile you own. Contradictions make you harder to name.
- Organization and Person schema. Explicit machine-readable identity with sameAs links tying your profiles together into one entity rather than several.
- Third-party corroboration. Mentions on sites you do not control. This is the part that cannot be faked and the part most agencies skip because it is slow.
- Content worth citing. Specific, verifiable, useful writing. Generic marketing copy is not quoted by anything, human or machine.
- Clear self-description. A page that plainly states what you do, for whom, and where. Obvious, and frequently absent.
- Crawler access. Deciding deliberately whether GPTBot, ClaudeBot, PerplexityBot and others may read your site. You cannot be cited by a system you have blocked.
- Monitoring. Asking the assistants the questions your buyers would ask, on a schedule, and recording the answers. The only real measurement available.
03 Crawlers
Who crawls what, and what it controls.
Frequently confused, and the confusion leads to genuinely damaging robots.txt decisions.
| Agent | What allowing or blocking it actually does |
|---|---|
| GPTBot | OpenAI’s training crawler. Blocking it keeps your content out of future model training. It does not remove you from ChatGPT’s live browsing. |
| OAI-SearchBot | Serves ChatGPT’s search and browsing. Blocking this does remove you from those answers. |
| Google-Extended | Controls use of your content for Gemini and Vertex AI. It does not affect AI Overviews in Google Search. |
| Googlebot | Ordinary search indexing — and the basis of AI Overviews. Blocking it removes you from Google Search entirely. |
| PerplexityBot | Perplexity’s crawler. Perplexity cites sources prominently, so this is usually worth allowing. |
| ClaudeBot | Anthropic’s crawler. Same reasoning as the others: allow unless you have a specific reason not to. |
Free GEO audit
What does ChatGPT say about your category?
I will ask ChatGPT, Gemini and Perplexity the questions your customers would ask, and send you the answers with the citations. It is usually the most useful hour in this whole field.
- Real answers from all three, screenshotted
- You keep the findings either way
- I will tell you what cannot be influenced
- Same-day reply, all 7 days
Free GEO audit
Get a free quote
Four details, and a real reply today.
04 Straight talk
What is genuinely influenceable.
The left column is worth paying for. The right column is worth being sceptical about.
You can influence
- How clearly you are described. On your site and everywhere else you appear.
- Whether sources agree. Consistency across profiles is the single most actionable factor.
- Whether anyone else mentions you. Slow, real, and the hardest thing for a competitor to copy.
- Whether crawlers can read you. A configuration decision entirely within your control.
You cannot control
- What the model already learnt. Training data is fixed until the next model. You cannot edit it.
- Whether you are named today. There is no submission, no bid, no dashboard.
- How the answer is phrased. The model composes. You are a source, not an editor.
- When it changes. Model updates move things unpredictably, in both directions.
05 Process
How a GEO engagement runs.
Audit the entity
Every place you appear online, and every way you are described. Inconsistencies listed and prioritised.
Fix and mark up
Descriptions made consistent, Organization and Person schema with sameAs, crawler access decided deliberately.
Earn corroboration
Mentions and citations from places you do not control. The slowest part and the one that actually differentiates.
Re-ask monthly
The same questions, recorded, compared. Including the months nothing moved, because that is also information.
— AI & Next-Gen Search
Related services.
These sit next to GEO and are usually bought with it. Same person doing the work in each case.
AEO Services
Be the answer, not result #7.
Read moreAnswer Engine Optimization
How answer engines actually pick.
Read moreGenerative Engine Optimization
The method, explained in full.
Read moreAI Search Optimization
AI Overviews, Perplexity, Copilot.
Read moreAI SEO Services
AI in the workflow, not the writing.
Read more— Questions
GEO — your questions.
Straight answers, including the ones that cost me work. If yours is not here, ask it — I reply the same day.
Generative engine optimisation: making your business one of the sources that AI assistants such as ChatGPT, Gemini, Perplexity and Copilot draw on and cite. The term comes from academic work published in 2023 and has since been adopted, somewhat loosely, by the marketing industry.
No. There is no submission process, no bid and no dashboard. Anyone guaranteeing this is describing something they do not control. What can be done is making you clearly described, consistently identified and independently corroborated — which is what these systems appear to favour.
It depends on your goal, and the distinction matters. Blocking training crawlers such as GPTBot keeps your content out of future models. Blocking retrieval crawlers such as OAI-SearchBot and PerplexityBot removes you from answers those tools give today. Most businesses that want to be recommended should allow the retrieval crawlers.
It shares a foundation — being crawlable, clear and credible helps everywhere. What genuinely differs is the emphasis on entity consistency across sites you do not control, and on third-party corroboration rather than on ranking. It is closer to public relations than classic SEO, which is why some agencies find it awkward to sell.
By asking the assistants a fixed set of questions on a schedule and recording what they say and cite. It is manual and imperfect and it is the honest method. Tools claiming a precise "AI visibility score" are producing a number, not a measurement.
For most small businesses it is not worth a large separate budget today. It is worth doing the cheap, durable parts — consistent descriptions, correct schema, deliberate crawler settings, one genuinely citable page — because those help conventional search anyway. I would rather tell you that than sell you a programme for a channel that is still forming.
Free consultation · No obligation
Find out what the assistants say about you.
One audit across ChatGPT, Gemini and Perplexity, with the answers and citations recorded. You keep it either way, and I will tell you plainly which parts are worth spending on.
- Real answers from all three, recorded
- I will tell you when it is not worth the budget
- No invented visibility scores
- You keep the findings either way
Start here
Send your enquiry
Name, number, email, what you need.