0 to 30,000+ combined monthly organic sessions across 10 clinic websites in 18 months.
Source: 18 months on a ten site clinic network, projects led by Muhammad Waqas.
Last updated November 2026
Generative engine optimization is the work that gets a business named and cited inside the answers ChatGPT, Perplexity, Google AI Overviews, Google AI Mode and Gemini generate.
What you buy here: one senior consultant owning your account, a specialist team I lead delivering it, a published method, and a monthly report built on prompts you agreed before the work started.
Everything below is the engagement rather than the theory: six pillars of work, five AI surfaces named in scope, a six step method with a document at the end of every step, and three dated client results. The longer answer to what generative engine optimization is opens the next section.
Free 30 minute strategy call on Google Meet, Monday to Friday, US Eastern business hours.
0 to 30,000+ combined monthly organic sessions across 10 clinic websites in 18 months.
Source: 18 months on a ten site clinic network, projects led by Muhammad Waqas.
Every outside number below names its publisher, its date and the limit that travels with it. Every client figure names the engagement behind it and how long that engagement ran.
Generative engine optimization, GEO, is the practice of making a business the source a generated answer is built from and names.
You are buying an engagement rather than a report: I own the strategy, the specialist team I lead ships the work, five AI surfaces are in scope by name, and every month measures a prompt list you approved first.
The mechanics are specific, which is why this is a separate discipline and not a tactic. An engine retrieves a handful of sources for a question, grounds its answer in them, picks which ones to name, and often fans one question out into several behind the scenes. The work is to be in that retrieved set, be nameable once you are in it, and be quoted often enough that your mention share on a prompt list moves. I am Muhammad Waqas, AI SEO consultant, and that pipeline from a question to a citation is what this page sells.
One naming note, because it saves confusion on a first call: I write generative engine optimization in full the first time it appears on any page, since the bare three letter token returns geological and geotechnical engineering results rather than this discipline.
A traditional result was clicked on 8% of visits where an AI summary appeared, against 15% of visits without one, and a link inside the summary was clicked on 1% of visits.
Pew Research Center, 22 July 2025: 900 US adults, 68,879 Google searches, browsing data from March 2025. Google only, by Pew's own stated limitation, and an association rather than a cause.
75.51% of surveyed buyers want the brand named in the answer even when it is not linked, against 14.29% who want the clickable citation.
Smarty Marketing, last updated 21 September 2026, whose published percentages reconcile to a base of 49 rather than the 300 the survey states.
More than half of small businesses block AI crawlers over content ownership and intellectual property, 57% still choose to restrict access, and 73% plan to revisit their policies within the year.
Clutch, 24 September 2025, a small business survey whose sample size, country and field dates are not disclosed and whose linked report is no longer available.
Three measured changes make this its own line of work rather than a paragraph inside an SEO report.
The result page answers before anyone clicks, the outcome buyers want has moved from the click to the mention, and most sites are restricting the very agents they want to be named by.
Pew's browsing study puts the gap at 8% of visits against 15%, with the link inside the summary taken on 1% (Pew Research Center, 22 July 2025, 900 US adults and 68,879 searches, browsing data from March 2025). A position that used to buy a visit now often buys a mention instead, and the mechanism behind that pattern is zero click search. The honest reading of Pew is an association on Google, not a proven cause, and it is still the clearest public measurement of what the summary does to a click.
The outcome buyers say they want has moved from the click to the mention: three quarters of the buyers surveyed for the figure in the stat row above put being named in the answer ahead of being the clickable source. Enterprise budgets have already followed the same preference: in Conductor's State of AEO / GEO in 2026 report, last updated 14 April 2026 across 250+ digital leaders, US enterprises with 500 or more employees allocated an average of 12% of digital budgets to this work in 2025 and 94% plan to increase it in 2026. Read that as enterprise behavior, because that is the population it measures.
Clutch's September 2025 report has more than half of small businesses blocking AI crawlers while 57% still choose to restrict access and 73% plan to revisit the policy inside a year (Clutch, 24 September 2025, sample and field dates not disclosed). Almost nobody arrived at that position on purpose. It is the cheapest thing in this discipline to get right and the thing most audits never open, which is why week one of every engagement starts there.
Six pillars carry the GEO service, ordered by the citation pipeline and not by SEO discipline: retrievability, entity resolution, answer shaped source pages, prompt and query coverage, third party corroboration, and citation measurement.
Each pillar states what it covers and what you receive, the five outcomes below are the measure, and the checklist names every real thing done.
Generative engine optimization is the named discipline inside a search engagement rather than a product bolted beside it, and the shape of yours is set on the free strategy call. For the wider program around it, see all my SEO services.
What the engagement produces:
Retrievability is the work that gets the engines in: robots.txt agent by agent, the CDN and firewall rules above it, bot management at the edge, whether your answer exists in the HTML before JavaScript runs, and the status codes on the pages you want quoted.
You receive an engine that reaches the pages you want cited.
The crawl budget record this has produced is in the case results below, and a technical SEO consultant engagement runs the same checks across a larger estate.
Entity resolution is the work that makes you one business an engine can name: the schema graph, Person and Organization nodes, Service and FAQPage markup, entity.json, llms.txt, one consistent name and role on every page, and the sameAs set behind them.
You receive one resolvable business rather than three spellings of one, declared identically on every page and in every file.
Answer shaped source pages are the pages an engine can quote without rewriting them: a liftable answer under every heading, the answer bolded rather than the keyword, definition blocks for the terms you own, question headings from live People Also Ask, and a sourced figure per claim.
You receive the pages written or rewritten so one sentence can be lifted as the answer.
Every page goes through my own 11 stage content pipeline, which holds an 8 out of 10 gate across six metrics and has run across 850+ pages. The peer reviewed paper that named this discipline, "GEO: Generative Engine Optimization" (KDD 2024), reports that adding citations, quotations and statistics to a source produced a 30 to 40% relative improvement on its own position adjusted visibility metric across a 10,000 query benchmark the authors built (Aggarwal et al., arXiv 2311.09735, v1 November 2023, revised June 2024), measured on that benchmark against GPT-3.5 rather than on live commercial search.
Prompt and query coverage is the inventory of buyer questions the work is built against: the prompt list itself, the commercial questions your buyers type, comparison and alternative prompts, query fan out behind one question, and one owning page per question.
You receive a map of the questions you intend to be the answer to, with the gaps marked and an owner assigned to each one.
Query fan out is the part most audits never open: one buyer question becomes several behind the scenes, and a business that answers only the first one drops out of the follow ups.
Third party corroboration is the off site half of this work: mentions in the sources engines already quote, unlinked mention reclamation, co citation beside the right companies, digital PR aimed at citable pages, and contributions under a real name.
You receive claims about your business that an engine can verify away from your own domain.
Citation measurement and reporting is the evidence that the rest of it landed: citations by prompt, mention share against competitors you choose, AI referred sessions separated from organic, your own Search Console and GA4 figures, and a monthly report that names what moved.
You receive a report you can audit by running the prompts yourself before the call.
What the work covers, pillar by pillar: six pillars and 65 named items, every one of them work I actually do rather than a capability list.
The pillars run in the order the citation pipeline runs, so the list reads the way the engagement is sequenced. Each pillar closes on one thing you can act on today.
GPTBotOAI-SearchBotPerplexityBotClaudeBotGoogle-ExtendedTip: the CDN decides before robots.txt does, because a request stopped at the edge never reaches the file. Check both, in that order.
Tip: give each entity one owning page, then let every other page mention it and point at that owner.
Tip: when no single sentence under a heading works as the answer, fix the heading first. The copy is rarely the real problem.
Tip: choose the prompts before anything is built. A baseline agreed after the work has started is not a baseline.
Tip: an engine reaches for sources it already trusts, so a mention on one of those pages is worth more than a link from a page nobody quotes.
Tip: if you cannot reproduce a citation by running the prompt yourself, it is not a result worth putting in a report.
Book a free 30 minute strategy call
Week one starts with the AI crawler access check on your own domain.
Five surfaces are in scope by name: ChatGPT, Perplexity, Google AI Overviews, Google AI Mode and Gemini.
Each is won by different work, and the table states what I do on it and what you see. Microsoft Copilot, Claude, Grok, Meta AI and the voice assistants are watched and reported on rather than sold as separate surfaces.
| Surface | What I do | What you see |
|---|---|---|
| ChatGPT | Open the search gate on purpose, give each buying question one sentence worth lifting, and earn mentions where the model already reads. In practice that is all how to rank in ChatGPT means | Your name in the answer, and your page behind it, when a buyer asks who does this |
| Perplexity | Keep the crawler welcome, put the answer in one findable place rather than three, and build sources worth citing. Practitioners file that work under Perplexity SEO | Your page in the numbered citation list printed next to the answer |
| Google AI Overviews | Keep the page indexed and snippet eligible, then make it the clearest statement of the definition and the figures a summary needs. That is page level work, which is where AI Overviews optimization lives | Your page listed as a supporting link beneath the summary on the queries that sell |
| Google AI Mode | Answer the question and its follow ups on one page, so a question that fans out behind the scenes still lands back on you. Practitioners call that Google AI Mode SEO | Your business still named at the third follow up, not only at the first question |
| Gemini | Make the entity resolvable in the knowledge graph, keep the Google-Extended decision a deliberate one, and maintain the files the assistants read first | Your brand described accurately when a buyer asks what the options in your category are |
Five surfaces are sold and the rest are watched, which keeps the scope honest: a surface goes in this table when I can say what I do on it and what you will see.
Eight things change in the work once the win condition is a citation rather than a position, and four competing names describe the same job from different angles.
The names come first, because buyers search all four and I need you to know which one you are reading.
| The name | What it means | Where it sits |
|---|---|---|
| GEO, generative engine optimization | Making a page the thing a generated answer is built from and names. 40.82% of surveyed buyers use this word (Smarty Marketing, 21 September 2026, base of 49 respondents) | This page, as the service you buy |
| AEO, answer engine optimization | Making a page the direct answer to a question so an engine can lift it whole. Only 6.12% of those buyers use the term, which is why I define it here rather than using it as a label (same source, same date and same base) | The same work under a different name, which answer engine optimization services covers in full |
| LLM SEO | Optimizing for the language models behind the assistants rather than for a results page. Not broken out separately in that source | The language model half of the same job, which this page covers as pillar 3 and pillar 4 rather than selling separately |
| AI SEO | The whole engagement across Google organic search and the AI surfaces together. 48.98% of those buyers call the work "SEO for AI" (same source, same date and same base) | AI SEO services for the combined program, and the concept itself in what AI SEO is |
On English Wikipedia, "Answer engine optimization" and "AI SEO" both redirect to "Generative engine optimization" (my own check, 9 November 2026), which is as short an answer as the question of GEO, SEO and AEO compared ever gets.
| What changes | Traditional SEO | Generative engine optimization |
|---|---|---|
| The unit of optimization | The keyword, and the page built for it | The entity, and the attributes an engine can state about it |
| The win condition | A ranking position on the result page | A named citation inside the generated answer |
| The reference layer | Links pointing at the page | The sources engines already quote, plus the knowledge graph |
| Access control | robots.txt written for Googlebot | Named agent lines: GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot and Google-Extended |
| The page shape | Long form built for dwell time | One liftable answer under every heading |
| The metric | Average position | Citation and mention share, prompt by prompt |
| The surfaces | One result page | Five sold by name, more watched |
| The reporting unit | A rank table | Named prompts you can run yourself |
Every row above implies a deliverable, which is why the method below ends each step in a document rather than a status update.
Six steps, in this order, each ending in something you hold: entity clarity, access, answer shaped pages, coverage, third party validation and measurement.
I publish the Entity to Answer method in full rather than describing it as proprietary, so you can read the sequence before you buy it.
The order is fixed because each step needs the one before it. Writing answers for an engine that cannot fetch the page is wasted work, and so is opening the gate to pages that resolve to nobody in particular.
<span class="lbl">What happens:</span> your business becomes one resolvable entity. The name, role and description are made identical on every page, the schema graph is rebuilt so each page references the same nodes, the knowledge panel is aligned where one exists, and the file that tells assistants which pages to read first is written. That alignment work is what practitioners mean by entity SEO, and llms.txt is one of the two files it produces.
Deliverable: the entity and schema map, with llms.txt and entity.json written.
<span class="lbl">What happens:</span> the gate gets opened on purpose. Your robots.txt is read line by line, the search gates are separated from the training gates, and the CDN and firewall rules above them are inspected, because an edge rule overrules anything robots.txt invites.
Deliverable: the AI crawler access report, across robots.txt, the CDN and the bot rules.
<span class="lbl">What happens:</span> every heading on the priority pages earns a complete answer inside its first 40 to 60 words, the entity gets named instead of referred to, each figure sits in the same sentence as its source, and the detail goes underneath. The headings themselves come from the People Also Ask questions that are live now.
Deliverable: the answer capsule and heading rewrite set for the pages you want cited.
<span class="lbl">What happens:</span> your buying questions are listed in the words a customer would use, one page is made the owner of each, the gaps get an owner too, and the prompts where a competitor is named and you are not become the publishing order.
Deliverable: the prompt and query coverage map, with one owning page per question.
<span class="lbl">What happens:</span> the claims on your site get corroborated somewhere else. Mentions are earned on the pages engines already quote in your category, unlinked mentions are reclaimed, and contributions go out under a real name. The detailed version of this step is how to get cited in AI answers, and the off site half of it is brand mentions for AI citations.
Deliverable: the corroboration plan, naming the sources to be named on and the mentions to reclaim.
<span class="lbl">What happens:</span> each month the agreed prompt list is re run the same way, beside your own Search Console and GA4 figures, and the report says which answers started naming you and which did not.
Deliverable: the monthly citation report, naming what moved and what did not.
Three artifacts land in month one, and each is a document you keep rather than a status update: the AI crawler access check, the entity and schema check, and the prompt baseline.
They are the first month versions of steps 2, 1 and 4, which is why week one already produces a finding on your own domain. Bought on its own, that first month is the AI visibility audit.
OAI-SearchBot, PerplexityBot, ClaudeBot, Google-Extended)In week one I read your robots.txt, the files beside it and your snippet directives, and I tell you agent by agent what each vendor may currently do with your pages.
The table below is what each of the five lines actually controls, and the distinction most sites miss is that only two of them are about being shown in an answer at all. The gate work itself is what practitioners call AI crawler access.
| Agent | Who runs it | What it controls |
|---|---|---|
OAI-SearchBot | OpenAI | Whether you are shown in ChatGPT search answers |
PerplexityBot | Perplexity | Whether you are surfaced and linked in Perplexity results |
GPTBot | OpenAI | Model training, which the documentation separates from ChatGPT search |
ClaudeBot | Anthropic | Model training |
Google-Extended | Gemini training and grounding, and Google states it does not affect inclusion or ranking in Google Search |
Every row comes from the vendor's own documentation, read 10 November 2026. The line that surprises people is the one that is not an agent at all: a page becomes eligible for Google AI Overviews and AI Mode by being indexed and eligible to be shown with a snippet (Google Search Central, read 10 November 2026), so what matters there is noindex, nosnippet, data-nosnippet and max-snippet rather than an allow line. You receive the file as it stands, a verdict per agent, and the exact replacement.
I read each schema block against what its page visibly shows, compare the naming from one page to the next, and take the sameAs set profile by profile against the accounts you confirm are yours. You receive the corrected schema graph, the entity map behind it, and a list of the claims your markup made that no page supported. Where the audit is the whole of the first engagement, the audit itself is the paid first step and the whole of the first month, and the classic side of the same ground is the technical and content audit inside my SEO services.
The prompt baseline is a dated written record of the buying questions your customers actually put to an assistant, and of what each engine answers today, taken before any work starts so every later month has something to be compared with.
One row per prompt, with six columns: the prompt in the words a customer would type, the engine it was run on, the date and time of the run, whether your business was named, who else was named, and what the answer leaned on where that engine shows it. Nothing else goes in it, because a column you cannot check by re running the prompt is not evidence.
The list is built from three places and then you decide: the questions your sales calls and your inbox already get, the questions your site already answers, and the questions your named competitors would want to win. You strike and add before it is frozen, and frozen means the wording stops changing, because a reworded prompt resets the comparison. It holds your brand name and the prompts you approved, run on public engines the way a member of the public would run them, and no customer data of any kind.
Ten checks on your own site, each one runnable in minutes, with the answer that means trouble written beside it.
They are on this page rather than behind an email form, because most of what I find in a paid audit starts with these ten and a buyer should be able to see that before hiring anybody.
| # | The check | What a bad answer looks like |
|---|---|---|
| 1 | Open your robots.txt and look for GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot and Google-Extended. A tool to check AI crawler access does the same read in one click | Any of the five disallowed while you are trying to be cited |
| 2 | Ask whoever runs your CDN whether AI user agents are blocked at the edge or in a bot rule | Nobody knows, which is the usual answer |
| 3 | View source on your most important page and search the HTML for the first sentence of its answer | The sentence is not there, because it only renders after JavaScript |
| 4 | Read the first two sentences under your top three headings | Neither sentence answers the heading |
| 5 | Search your own brand name and role as one phrase and read what comes back | Another business or another person with your name |
| 6 | Check that your name, role and description match on every page and every off site profile | Three versions of one business |
| 7 | Look for llms.txt and entity.json at your root | Neither exists |
| 8 | Run your five most commercial prompts in ChatGPT and Perplexity and note who is named | Competitors named, you not named |
| 9 | Check whether a date appears on the pages that make claims | No date on anything |
| 10 | Count how many of your pages answer the same question | Two or more, which is why neither is the one quoted |
Anyone can work through those ten in an afternoon, and the paid audit starts where they end.
Book a free 30 minute strategy call
Ten checks you can run yourself. The paid audit starts where they end.
Six findings come up in almost every GEO audit I run, and each one costs the site something specific: blocked crawlers, an answer that only exists after JavaScript, inconsistent naming, two pages on one question, schema the page does not support, and no baseline.
These are what the ten checks above turn up, in the order the engagement produces them.
I measure generative engine optimization as the agreed prompt list re run on the same cadence, set beside the organic figures out of your own Search Console and GA4, once a month, on a call with the people who did the work.
Six measures carry that report, and three familiar numbers never appear in it as results.
What is measured and reported:
Each month's run sits beside the frozen baseline, same wording, same engines, same point in the month, so the change is visible without anybody's interpretation.
What is never counted here as a result: domain rating, domain authority and brand only traffic.
The first two are third party scores no engine consults when it picks a source, and the third is demand you already had. Picking the prompts, running them and dating the result is also the whole of how to measure AI visibility, and the bar every figure on this site clears is the one printed beside it: publisher, date and limit. The client figures carry the same bar in their own terms, the work and the period behind each one, and more of them sit in my AI SEO case studies.
By the end of month 1 you hold the written baseline and an applied fix list, months 2 to 3 bring the first citations on named prompts, months 4 to 6 move mention share and head terms, and months 6 to 12 compound.
Four bands, each naming the work inside it and what you hold when it closes.
28% of search specialists advise a lead time of at least three to four months, and half of them say businesses underestimate how long results take (Fiverr index via Search Engine Journal, 22 September 2026, a freelancer survey whose total sample and field dates were not disclosed). That matches what I see, and how long SEO takes on your own site depends on its state and the competition in front of it.
| Period | What is done | What you have by the end of it |
|---|---|---|
| Month 1 | Entity and schema map built, AI crawler access fixed, audit findings ranked, prompt baseline frozen | The written baseline, the applied fix list, and a publishing order you approved |
| Months 2 to 3 | Rewrites on the pages that already rank, retrievability fixes shipped, the first new source pages live | First movement on long tail questions and the first citations on named prompts |
| Months 4 to 6 | Coverage build out across the prompt map, internal entity linking, corroboration placements landing | Mention share climbing and head terms rising on the prompts in your own list |
| Months 6 to 12 | Compounding: more coverage, more corroboration, migration or crawl work where the site needs it | Commercial questions held across the surfaces, and inquiries the report can attribute |
With the right budget and the time the work takes, the plan above outranks the competition and gets you named in the answers around it.
Six situations put a business into generative engine optimization services: assistants naming your competitors, traffic falling while rankings hold, being smaller than the brands named, a migration or hijack that broke the pages you want quoted, an instruction to be visible in AI search this quarter, and an in house team that needs leading.
Each has its first move.
A buyer asks for options in your category and your name is absent. First move: the entity and schema check, because an engine cannot name a business it cannot resolve.
The positions are intact and the clicks are not. First move: measure what the summary is taking on your commercial questions, then own the definition and the figures it needs.
Smaller businesses compete here on specificity rather than on budget. First move: the narrow prompts where you are genuinely the clearest answer, because that is where a precise source beats a big one.
Redirects, duplicate URLs or injected pages are standing between an engine and your answer. First move: the retrievability pass. Hijack recovery and large site migration are the same work as this, and the scale of both is in the career scope further down.
The instruction arrived without a method attached. First move: the frozen prompt baseline, so the quarter ends with a measurement rather than an opinion.
You have the people and want the discipline. First move: the published method as the operating standard, with me leading the work and training your team against it, which is how a team learns to run it without me.
Four routes get compared here on six buyer criteria: this engagement, a GEO or AI visibility tool as a category, a traditional SEO agency, and an in house hire.
Six rows decide it: who picks the prompts, who changes the code after a finding, who owns the entity graph, who earns the mentions, and what you keep.
| What you are comparing | This engagement | A GEO or AI visibility tool | A traditional SEO agency | An in house hire |
|---|---|---|---|---|
| Who chooses the prompts | A named list agreed with you and frozen before the work starts | Whatever the product tracks by default | Whatever the agency reports on | Whoever has time that month |
| Who does the work after the finding | The work itself, by the team that found it. SEO consulting services describes that senior ownership in full | The finding, on a dashboard, with the doing left to you | The classic SEO work, with the AI surfaces often outside scope | One person, at one person's capacity |
| Crawler access and rendering fixes | Changed: robots.txt, the CDN rule and the HTML | Reported where detected | Changed where the scope covers it | Changed once a developer is free |
| Entity and schema ownership across the whole site | Page by page, as one graph, for as long as the engagement runs | Monitored rather than owned | Per campaign or per page shipped | Owned, alongside everything else on the list |
| Third party corroboration | Mentions earned on the sources engines already quote | Outside what software can do for you | Where link or PR work is in scope | Limited by one person's network |
| What you hold at the end | The published method, the frozen prompt baseline and a monthly report you can re run yourself | Access, for as long as the subscription runs | The reports delivered during the term | The experience, and the gap when they leave |
One senior consultant owns your GEO account and leads the specialist team that delivers it, and the years and the team scale behind that strategy are stated at the top of this page.
I set the plan and sit on your monthly call. The delivery runs under me, across content, technical work, healthcare, food brands, design, development and digital PR.
What that career covers, stated in full rather than narrowed to one example: around 50 hijacked websites recovered; websites with thousands of pages migrated without lost equity; clinic and healthcare sites grown to their maximum organic traffic and conversions; food, restaurant and catering brands grown; and parking management and ecommerce websites grown.
What stays with me on every GEO engagement, and what ships under me:
The content pipeline behind the source pages is work I led as well, and the gate it holds is named in the deliverables above. The career behind the strategy sits on about Muhammad Waqas, and the same senior ownership across a whole search function rather than GEO alone is SEO consulting services.
Certifications:
Three results from projects I led, each with its problem, the work and the figure: a ten site clinic network, crawl budget across a large estate, and a migration of a site running into thousands of pages.
Each period is the one the work actually ran for.
Book a free 30 minute strategy call
Those three figures are the record behind this engagement. There are more of them in my AI SEO case studies.
An AI answer already sits on the GEO buying searches your buyers make, and I have the dated reads to show it.
Two measurements of my own, plus one dated line on the criterion this market's own weighted ranking puts first. Both are reads of the live result pages here, dated, and saved as files I can open again.
An AI Overview sat on generative engine optimization services on both of my reads, on 9 and 10 November 2026, and on three of the four other GEO buying searches I read on 10 November: generative engine optimization agency, generative engine optimization company and geo seo services. It was absent on the bare geo agency, which returned images, a local pack and paid results instead, because that token is ambiguous. Five queries, read in the United States. My reads record that an AI Overview is present, not its text and not the sources it cited.
The same result page moved in the 23 hours between those two reads, on 9 and 10 November 2026. Five of the nine results changed and four held: webfx.com left position 2 and semrush.com position 3, and bluecompass.com entered at 1. Two reads are a volatility observation rather than a trend, and they are why I re read this market every 14 days.
The one weighted ranking method published in this market puts qualified leads first, at 41.1% (First Page Sage, read 10 November 2026).
I serve three industries here, and each one carries either a figure or a named client: clinic groups in healthcare, restaurant and food brands, and the technical work behind large sites and migrations.
Agencies list the sectors they would like to work in. These are the three I can evidence.
Multi location clinic groups, where every location competes for its own searches and the group still has to be the named source in an answer without its branch sites eating each other. The record behind that is the ten site network in the case results above, with its figure, its period and its source on the card. At that scale, healthcare SEO services get planned at network level rather than site by site.
Brands whose menus, locations and ordering pages have to be findable in Google and quotable inside an answer, in a category where the decision is made in minutes. The work here covers location pages that do not duplicate one another, menu and dish entities an engine can parse, and the profile presence the brand already has. Captain Zyke is one of the brands in this sector I have done SEO for, and work of that shape goes by the name restaurant SEO.
Large sites where the problem is not content but access: what gets crawled, what stays indexed, what a migration broke, and whether the answer is in the HTML at all. What stands behind it is the crawl budget and migration cards in the results above. The work here is the redirect map drawn page by page before anything moves, the parameter and version URLs lifted out of the crawl path, and server side rendering where the answer only existed after JavaScript ran.
Six reasons to hire me for this work, and each is either a proof item or a commitment I publish, never an adjective.
Ten years, senior ownership of the team I lead, a method in the open, a source beside every figure, the three numbers I refuse to report, and client results with figures. All six are evidenced above.
Ten years of client accounts, and the evidence for it is three dated results further up this page instead of a line in a tagline. Each one states what was wrong, what I did and the number that moved, over a stated period.
The strategy, the frozen prompt list and the monthly call are mine. The specialist team I lead ships against that plan, inside scopes I set and review. Decisions get seniority and delivery gets capacity, which is a different thing from one person doing both.
The six steps are the real sequence, in the order I run them, with the document each step produces written beside it. None of it waits behind a signature and none of it is called proprietary.
Each outside figure on this page prints its publisher, its date and the limit that travels with it, and each client figure prints the work and the period behind it, including the two result page reads of my own with the dates I read them. Nothing here asks to be taken on trust.
Three numbers never enter a report I write: domain rating, domain authority and brand only traffic. You read that commitment here, before an engagement starts, rather than discussing it in month four.
Three results, each with its figure, its period and its industry, so you can find the one that sits closest to your own business and judge the number on its own terms.
Twelve generative engine optimization questions, each answered completely in its first sentence.
Every one is a question buyers really ask: live People Also Ask captured on my own reads of this market, a question two or more of the pages ranking for this query answer, or one from my own inquiries.
A generative engine optimization consultant is the person who makes a business the source a generated answer names, and who owns that outcome rather than reporting on it. On my engagements that means the strategy, the prompt list and the monthly call, much like what an AI SEO consultant does.
Six things, in order: entity clarity, crawler access, answer shaped source pages, prompt and query coverage, third party corroboration, and monthly citation reporting. The method above is the long answer, with the deliverable named at every step, and it overlaps with what an AI SEO consultant does.
No: GEO decides whether you are named in the answer, and it runs on the indexed, crawlable pages classic SEO builds. Both get bought together on every engagement I run. The eight row table above states what changes, and the longer version sits in GEO, SEO and AEO compared.
The scope and the cost are set on the free strategy call, and an engagement is invoiced in USD. Worth is answered by what lands: the baseline, the fix list applied, the source pages, the citations and the monthly report. See how scope and terms are agreed and SEO ROI.
It is evolving: the demand did not move, the place the answer appears did. Buyers still research before they buy, and the work now covers retrievability, entity resolution and prompt coverage on top of rankings. That is the short answer to is SEO dead.
Citations by prompt, mention share against competitors you chose, prompt level presence you can re run yourself, AI referred sessions separated from organic, and your own Search Console and GA4 numbers. Reported monthly against a frozen baseline. Domain rating, domain authority and brand only traffic are never reported as results.
Month 1 ends with the written baseline and the fix list applied, months 2 to 3 produce the first citations on named prompts, months 4 to 6 move mention share and head terms, and the gains compound from month 6. The table above names what you hold at each stage.
Yes as a foundation: Person, Organization, Service, FAQPage and BreadcrumbList make the facts on a page machine readable, and schema by itself does not buy a citation. It earns its place when the markup matches what the page visibly says. Markup describing things a reader cannot see works against you.
In ChatGPT: let OAI-SearchBot fetch you, put a liftable answer on the page, and earn mentions on the sources the model already quotes. In Google AI Overviews: keep the page indexed and eligible to be shown with a snippet, then own the definition and the figures the summary needs.
Both, and you choose which: GEO runs as the named discipline inside a search engagement, or as its own engagement where the SEO foundations are already sound. The shape of yours is set on the free strategy call, after I have read what your site does today.
Yes, on specificity rather than budget: an engine picks the clearest source for a question, not the biggest brand in the category. A smaller business that owns the narrow buying questions in its field gets named on them. That is why the prompt list comes before the work.
Yes: I lead the work and train your team against the published method, which is how I run the specialist team I lead. You get the standard written down, the reviews run against it, and the prompt baseline kept in your account rather than mine.
Three routes exist into this work, and the right one depends on how much of your search program you want me to own: generative engine optimization on its own, the combined program across Google organic search and the AI surfaces, or senior ownership of the whole search function.
Each is a separate engagement, each set on the same call.
All three run the published method, the frozen prompt baseline and the same monthly report. The dated results behind them are in my AI SEO case studies, and the career behind the strategy is on about Muhammad Waqas.
Bring your domain to the call and leave it knowing three things: whether the engines can read your site today, the first thing I would change, and why that one comes before the rest.
Thirty minutes on Google Meet, with the consultant who would own your account.
Book a free 30 minute strategy call
The specialist team behind the work is mine, and every figure above carries the engagement it came from. If writing suits you better than a call, contact me with the details instead.
Free 30 minute strategy call on Google Meet, Monday to Friday, US Eastern business hours.