SERVICE / 10
AI SEO / Generative Engine Optimisation
Search visibility is expanding beyond ten blue links. AI Overviews, AI Mode and answer engines like ChatGPT and Perplexity now sit between your content and your customer, and they cite a small number of sources. I help companies understand what actually influences those citations, build the content and technical foundations that earn them, and measure the result without pretending the measurement is more precise than it is.
On this page
There is a lot of confident advice about "ranking in ChatGPT" and very little evidence behind most of it. What I can say with reasonable confidence, from my own tests and what the engines publish, is that AI answer systems draw heavily on the signals that have always mattered: pages that are accessible, clearly about one thing, authoritative, and written so a machine can extract a direct answer. The work is not new. The emphasis has shifted and the measurement is harder. That is the honest starting point.
What AI SEO and generative engine optimisation mean
AI SEO, also called generative engine optimisation (GEO), is the practice of making a brand and its content visible inside AI-generated answers: Google's AI Overviews and AI Mode, ChatGPT's search-backed responses, Perplexity, Copilot and similar systems. These retrieve a set of web pages, synthesise an answer and cite some sources. Visibility means being one of those sources, or being named in the answer itself.
Two mechanisms matter. Retrieval-backed systems run searches against a conventional index, pick pages and build an answer from them; being crawlable and relevant is the entry ticket. The model's own knowledge depends on how consistently the brand is described across the web over time. Neither is something a plugin can fix.
- Query
- Retrieval
- Selection
- Synthesis
- Citation
- Click or brand recall
Who this is for
SaaS and B2B companies
Where buyers ask AI assistants for recommendations and "best tools for X", and absence from the answer means absence from the shortlist. This connects directly to my SaaS SEO work.
Publishers and content-led businesses
Sites seeing traffic erode on informational queries that AI Overviews now answer, and needing a strategy for what comes next.
Brands with a category to define
Companies that want to be the entity an AI names when someone asks about a product category, especially where the category is new.
Problems this solves
Traffic falling on queries you still rank for
You are on page one, but an AI Overview answers above you and cites someone else. The fix is to become the cited source, or shift effort to queries where the click still happens.
Not appearing in AI recommendations
Competitors are named when someone asks an assistant about your category; you are not. This is usually an entity clarity and third-party corroboration problem, not a content volume problem.
Content that cannot be extracted
Good information buried in long introductions or vague headings, or rendered in a way AI crawlers do not execute.
No way to know if any of it is working
Leadership asks "are we in ChatGPT?" and nobody has a defensible answer. I set up measurement that is repeatable and clear about its limits.
What actually influences citation
This is what I have found to be true in practice, in my AI search citation test and in client work. I present it as a set of factors with varying confidence, because that is what it is. The AI SEO guide goes into each in more depth.
Technical accessibility
If a crawler cannot fetch and read the page, nothing else matters: allow the relevant bots (or decide deliberately not to), server-render content, keep HTML clean. I test access per bot rather than assuming; my JavaScript rendering test is relevant here.
Conventional search visibility
Retrieval-backed systems lean on search indexes, so pages that already rank for the underlying query are far more likely to be retrieved. AI search does not replace fundamentals; I argue this in AI search and fundamentals. Strong technical SEO and topical authority are prerequisites.
Extractable answers
Pages where the direct answer to a specific question sits under a specific heading, in plain language, early in the section, are easier to cite. Definitions, step lists, comparison tables and labelled data all help. It is the same discipline as writing for featured snippets, with more at stake.
Clear entities
Brand, product and people described consistently, with structured data (Organization, Product, Person), an about page that says plainly what the company does, and the same facts on the external profiles that models and retrieval systems consult. Semantic SEO thinking applies directly.
Original data and first-hand experience
Answer engines prefer sources that say something no one else does: proprietary research, benchmarks, test results, first-hand accounts. Rewritten summaries of other people's content are the least likely to be cited because the engine can use the original.
Brand mentions and corroboration
Being named on sites engines trust (industry publications, review platforms, community discussions) appears to influence whether a model recommends a brand by name. This is digital PR work: slower and less controllable than on-page changes, which is exactly why it is hard to fake.
AI search changes which pages get the click. It has not changed what makes a page worth citing.
What is hype
| Claim | My view |
|---|---|
| "We can get you ranked in ChatGPT" | There is no ranking to get. Citations vary per query and per session. Anyone promising placement is promising something they cannot control. |
| "Add an llms.txt file and AI engines will prioritise you" | An emerging convention with no confirmed effect on citation at the time of writing. Harmless to add, not a strategy. |
| "GEO replaces SEO" | Retrieval-backed answers depend on search indexes. The fundamentals are the prerequisite. The emphasis shifts; the work does not disappear. |
| "Write content in a special format for AI" | Clear structure helps. A separate "AI-formatted" version of your pages does not, and risks creating duplication. |
| "AI visibility tools give you an accurate share of voice" | They sample prompts and report what they see: useful for direction, useless for precision. Treat the numbers as trends. |
Measuring AI visibility honestly
Measurement here is imperfect and I say so up front. I set up a repeatable process rather than a dashboard that implies false precision.
- A fixed prompt set. A panel of the questions your buyers actually ask, run on a schedule across the engines that matter for your market, recording whether you are cited or named and who else is.
- Search Console signals. Google folds AI Overview and AI Mode impressions and clicks into normal search performance data. I track the queries where they appear and watch the click-through changes they cause.
- Referral traffic from AI sources. ChatGPT, Perplexity and others send identifiable referrals: small volumes, often high quality. I report them separately.
- Brand search and direct traffic. Being recommended by an assistant often produces a brand search later rather than a click now. Changes in brand query volume are a lagging but real signal.
What you get
AI visibility baseline
The prompt panel, your current citation and mention status across engines, competitor presence, and an honest account of what the baseline can and cannot tell you.
Technical access audit
Bot access rules, rendering per crawler, CDN and firewall behaviour, structured data, and a recommendation on which crawlers to allow and why.
Entity and brand consistency plan
The canonical description of the company, products and people; structured data implementation; and a list of external profiles to correct or create.
Content restructuring and original-data plan
Which existing pages to restructure for extractability, which questions deserve direct-answer sections, and what original research or data your business is uniquely able to publish. This builds on my content SEO approach.
Corroboration and mention strategy
The publications, platforms and communities where mentions would matter for your category, and a realistic plan to earn them.
Measurement process
The scheduled prompt panel, Search Console and analytics segmentation, and a quarterly report written in plain language with its limitations stated.
How the engagement runs
Baseline and fundamentals check
I establish the prompt panel and baseline while auditing the conventional search foundations. If the site has crawl, index or authority problems, those come first; AI visibility work on a weak foundation does not hold.
Access and entity fixes
Bot access decisions, rendering fixes, structured data and the brand consistency clean-up. These are the quickest wins and the most frequently neglected.
Content restructuring
Rewriting key pages so each section answers one question directly, adding definitions, comparison tables and data where they belong, and tightening the content that already ranks.
Original data and corroboration
Producing the research or first-hand material that gives engines a reason to cite you specifically, and earning the external mentions that make the entity credible.
Measure and adjust
Quarterly review of the panel, the search data and referrals. Engines change; the approach adjusts with evidence rather than with each new announcement.
Related work
The AI search citation test documents my method for testing which page characteristics correlate with citation. Recent engagements with Yogesh Pant (Digiliance), Anil Sharma (Sentiment Furniture Systems), and Yogesh Tulani (Saniiq) demonstrate how established search foundations translate directly into citations within AI-generated answers across ChatGPT, Google AI Overviews, and Perplexity.
Much of this work is with technology companies in Bangalore and Hyderabad, where buyers are early adopters of AI assistants for research.
Frequently asked questions
Is AI SEO different from regular SEO?
It is an extension, not a replacement. The retrieval systems behind AI answers rely on search indexes and the same quality signals. What changes is the emphasis: extractable structure, entity clarity, original data and corroboration matter more; chasing volume on generic informational content matters less. A site with weak fundamentals will not be fixed by AI-specific tactics.
Should we block AI crawlers?
It is a business decision, and it should be made deliberately. Blocking training crawlers while allowing search and retrieval crawlers is a reasonable position for many businesses. Blocking everything means you will not be cited. Allowing everything means your content may be used in ways you did not intend. I lay out the trade-offs per bot and implement whatever you decide.
How do we measure whether we are showing up in ChatGPT or AI Overviews?
Through a fixed, scheduled prompt panel across engines, Search Console data for the queries where AI Overviews appear, referral traffic from AI sources, and brand search trends. None of these is complete on its own. Together they show direction reliably, which is what you need to make decisions.
Will AI Overviews kill our organic traffic?
For some informational queries, clicks do fall when an Overview appears, and that is already visible in many sites' data. For transactional, comparison and brand queries the effect is smaller, and being cited within the Overview can offset the loss. The strategic response is to shift the content mix toward the queries where clicks still happen and to become the cited source on the rest.
How long before we see any change?
Technical access fixes and content restructuring can affect citations within weeks, because the engines re-retrieve frequently. Entity consistency and corroboration take months. Because citation is non-deterministic, I look for trends across the prompt panel over quarters rather than week-to-week movements, and I would be cautious of anyone who reports faster.
Start with what is testable
If you want a clear picture of where your brand stands in AI answers and a plan built on the parts that can be tested, let's talk. I will tell you what I know, what I am unsure about, and what I would not spend money on.
Work with me
Have an SEO problem?
Tell me what you're working on.
Start a Conversation Explore SEO Resources