By Christoph Olivier, Founder, CO Consulting. Fractional CMO who has managed millions of dollars in ad spend for 7- and 8-figure businesses. Last reviewed: October 2026.
Rank on AI is my AI search optimization service. I make your brand easier for ChatGPT, Claude, Perplexity, Gemini, Microsoft Copilot and Google AI Overviews to find, trust and cite when buyers ask about your category. Each engine pulls sources from a different index, so the work covers crawler access, entity data, answer-first pages, original data and third-party mentions together. Nobody can guarantee a citation, and I do not.
- Who it is for: established firms whose buyers research before they call, especially in tax, M&A, consulting, real estate investing, capital raising and financial services.
- What you get: a measured AI visibility baseline, fixes across all major engines, rebuilt priority pages, an original data asset and a monthly share-of-voice report.
- How it is priced: scoped per engagement after a call, because the work depends on what your site and brand already have.
Book a consultation to find out whether AI search is the right lever for you this year.
What is AI search optimization?
AI search optimization, also called generative engine optimization (GEO) or answer engine optimization (AEO), is the work of getting your brand quoted inside AI-generated answers rather than only listed among blue links. The goal is to be the cited source or the named recommendation at the moment a buyer asks an assistant who to hire, which option fits, or what something should cost.
It matters now because the click is moving. The numbers below are the ones I would put in front of a board, each from a named third party:
| Finding | Source |
|---|---|
| Google users clicked a traditional search result on 8% of visits when an AI summary appeared, versus 15% when it did not. | Pew Research Center, July 2025 |
| AI Overviews were associated with a 34.5% lower click-through rate for the top-ranking page, across 300,000 keywords. | Ahrefs, April 2025 |
| AI Overviews reach more than 2 billion monthly users; AI Mode passed 100 million monthly active users in the US and India. | Alphabet Q2 earnings remarks, July 2025 |
| ChatGPT passed 800 million weekly active users. | OpenAI DevDay via TechCrunch, October 2025 |
| The average AI search visitor was worth 4.4 times the average traditional organic visitor, based on conversion rate. | Semrush, June 2025 |
Fewer clicks reach the open web, and the ones that do arrive further along in their decision. If your brand is missing from the answer, you often never see the lost lead in your analytics at all.
What is included in the Rank on AI engagement
Every engagement includes seven deliverables: an AI visibility baseline, a crawler and index access audit, an entity and schema layer, answer-first rebuilds of your priority pages, one original data asset, an earned-mention plan, and measurement. The mix of effort shifts by client, but nothing on this list is optional, because each one covers a different stage of how engines choose sources.
| Deliverable | What you get | Why AI engines care |
|---|---|---|
| 1. AI visibility baseline | A fixed set of buyer prompts run across ChatGPT, Claude, Perplexity, Gemini, Copilot and Google AI Overviews, logging who gets named and cited today. | You cannot improve share of voice you have not measured. The baseline also shows which competitors and third-party sites the engines already trust. |
| 2. Crawler and index access audit | Checks for Googlebot, Bingbot, OAI-SearchBot, Claude-SearchBot and PerplexityBot in robots.txt, firewall and CDN rules; Bing Webmaster Tools setup; indexation fixes. | An engine cannot cite a page its crawler is blocked from or its index does not hold. A firewall rule can do this silently. |
| 3. Entity and schema layer | Consistent name, address and author data across the web; Organization, Person and Service schema linked through sameAs to your real profiles. | Engines need to resolve your brand to one clear entity before they can recommend it with confidence. |
| 4. Answer-first page rebuilds | Your highest-value service and question pages rewritten so each section opens with a direct, quotable answer backed by sourced numbers and tables. | Retrieval works at the passage level. A clean 40 to 75 word answer under a clear heading is easier to lift and cite than a long preamble. |
| 5. One original data asset | A research piece built from primary data in your niche, published with method notes, like our own AI search readiness study. | Engines and journalists cite the original source of a number. Owning a statistic gives other sites a reason to mention you. |
| 6. Earned-mention plan | A target list of the publications, directories, review sites and community threads the engines already cite for your prompts, plus a plan to appear on them. | Off-site brand mentions correlate with AI visibility more strongly than backlinks do (see the Ahrefs data below). |
| 7. Measurement and reporting | A GA4 channel group for AI referrals, a monthly re-run of the baseline prompts, and a one-page share-of-voice report. | AI traffic is partly invisible in default analytics. Without a setup, real progress looks like nothing happened. |
How each AI engine chooses its sources
Each AI engine retrieves from a different place. ChatGPT search draws on third-party providers such as Microsoft Bing plus OpenAI’s own crawler; Claude’s web search has been traced to Brave Search; Perplexity runs its own index; Gemini and Google AI Overviews use Google’s index; Copilot uses Bing. Ranking well on Google alone does not cover the other four.
| Engine | Where it finds sources | Crawler to allow | What tends to earn the citation |
|---|---|---|---|
| ChatGPT search | Third-party search providers, with Microsoft Bing the widely reported one, plus content from OpenAI’s partners (OpenAI via TechCrunch, Oct 2024). | OAI-SearchBot. OpenAI states that sites opted out of it will not be shown in ChatGPT search answers (OpenAI crawler docs). | Bing indexation and off-site mentions. Semrush found nearly 90% of pages ChatGPT cited ranked outside Google’s top 20 for related queries (Semrush, 2025). |
| Claude | Brave Search, added to Anthropic’s subprocessor list when web search launched (TechCrunch, March 2025). | Claude-SearchBot and Claude-User. Blocking ClaudeBot only opts out of training (Anthropic support). | Presence in Brave’s index, primary data, and clear passages that answer the question directly. |
| Perplexity | Its own index of hundreds of billions of web pages (Perplexity, Sept 2025). | PerplexityBot, which surfaces and links sites and is not used for model training (Perplexity docs). | Citation-first answers: every response lists sources, so fresh, specific, well-structured pages get picked. |
| Google AI Overviews and AI Mode | Google’s index. A page must be indexed and eligible to show with a snippet; Google says no special optimization is needed. Both may use “query fan-out” across subtopics (Google Search Central, Dec 2025). | Googlebot. | Classic SEO fundamentals plus pages that answer the related sub-questions fan-out generates. |
| Gemini | Grounding with Google Search, which connects Gemini to real-time web content and returns cited URLs (Google AI for Developers). | Googlebot. | The same Google index signals, plus a clean entity footprint. |
| Microsoft Copilot | Bing. Microsoft states that Bing and Copilot search rely on the same crawling, indexing and ranking foundation (Bing Webmaster Guidelines). | Bingbot. | Bing indexation, which also feeds ChatGPT, so one fix serves two engines. |
The practical takeaway: a single “AI SEO” checklist misses at least one of these pipelines. The engine-by-engine detail lives in my guides on how to rank on ChatGPT, how to get cited by Claude and how to rank on Perplexity.
The four stages every AI answer passes through
Before an AI engine cites you, your page has to clear four stages: a crawler fetches it, a search index stores it, the engine retrieves it when it turns a prompt into searches, and the model picks it as a source it trusts. A failure at any stage ends the chain, which is why the engagement works on all four.
Many “AI SEO” offers only touch stage 3, rewriting pages. The problems that get missed usually sit in stage 1 (a firewall or CDN rule quietly blocking a search crawler) and stage 4 (a brand barely mentioned anywhere the engines look).
Why off-site mentions matter more than links for AI visibility
Off-site brand mentions are the strongest measured predictor of AI visibility so far. In Ahrefs’ study of 75,000 brands, branded web mentions correlated with appearing in Google AI Overviews at 0.664, against 0.218 for backlink count and 0.326 for Domain Rating. That is why one deliverable is an earned-mention plan, not a link campaign.
| Factor (Spearman correlation with AI Overview brand mentions) | Correlation |
|---|---|
| Branded web mentions | 0.664 |
| Branded anchors | 0.527 |
| Branded search volume | 0.392 |
| Domain Rating | 0.326 |
| Referring domains | 0.295 |
| Number of backlinks | 0.218 |
Source: Ahrefs, “An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)”, May 2025. Correlation is not causation, and this study covers Google AI Overviews only. On the content side, the Princeton and Georgia Tech team that coined the term GEO reported that optimization methods can raise visibility in generative engine responses by up to 40%, with results varying by domain (Aggarwal et al., KDD 2024).
How the engagement runs, step by step
The engagement runs in six phases: baseline, access, entity, pages, data and mentions, then ongoing measurement. The early phases are fast technical fixes that can show up as soon as engines recrawl. The later phases, earned mentions especially, compound over months. Timing depends on your starting authority, so I set dates on the scoping call rather than here.
- Scoping call. We check whether your foundation can support AI visibility yet. If it cannot, I say so and point you at what to fix first.
- Baseline. I agree a prompt set with you (the questions your buyers actually ask), run it across every engine and record who is named and cited.
- Access and indexation. Crawler rules, firewall and CDN settings, Bing Webmaster Tools, sitemaps and index coverage get fixed so every engine can reach your pages.
- Entity and schema. Your name, people and services are made consistent across your site and profiles, then marked up honestly to match the visible page.
- Pages and data. Priority pages are rebuilt answer-first, and one original data asset is researched and published with sources and method notes.
- Mentions and measurement. The earned-mention plan goes live, the baseline prompts are re-run monthly, and you get a share-of-voice report you can read in five minutes.
Book a scoping call and I will tell you which phase your site would start in.
Is Rank on AI the right fit for your business?
This service fits established firms with real expertise, some organic presence and buyers who research before contacting anyone. It is a poor fit for brand-new sites, for anyone expecting guaranteed placements next month, and for businesses without basic analytics. If two or more “not a fit” rows describe you, build the foundation first.
| Situation | Fit | Why |
|---|---|---|
| Established site with real content and some organic authority | Right fit | Engines already have something to retrieve. The work makes existing authority easier to read and cite. |
| Considered-purchase or B2B firm where buyers research before calling | Right fit | Those buyers increasingly ask assistants for shortlists. Being named there reaches them early. |
| Firm with proprietary data, cases or expertise nobody else publishes | Right fit | Original numbers are what engines and journalists cite. You have the raw material for a data asset. |
| Brand-new site with no content, reviews or outside mentions | Not a fit yet | There is little for an engine to retrieve or trust. Classic SEO and content come first. |
| You want guaranteed citations or rankings by a fixed date | Not a fit | The platforms do not sell citation slots. I only work with conditional outcomes. |
| No analytics, inconsistent business details across the web | Not a fit yet | Fix the fundamentals first, or you will not be able to tell whether anything worked. |
How results are measured
Results are measured three ways: AI share of voice (how often you are named or cited across a fixed prompt set, compared with competitors), citation detail (which pages get cited, by which engine), and AI referral traffic with its conversions in GA4. Referral numbers undercount real influence, so I read all three together rather than trusting one figure.
| Metric | What it tells you | How it is tracked |
|---|---|---|
| AI share of voice | How often your brand appears in answers to your buyers’ prompts, versus named competitors. | Monthly re-run of the baseline prompt set across all six engines. |
| Citations by page and engine | Which of your URLs each engine cites, and which third-party pages it cites instead of you. | Logged from the same prompt runs; feeds the next round of page and mention work. |
| AI referral sessions and conversions | Visitors who clicked through from an assistant, and what they did next. | A dedicated GA4 channel group for AI referrers, reported alongside organic. |
| Branded search demand | Whether people who saw you in an answer later search for you by name. | Google Search Console and Bing Webmaster Tools branded query trends. |
A citation inside an answer often produces no click, and some assistant traffic arrives without referrer data and lands as “Direct” in analytics. That is why share of voice leads the report and traffic follows it.
What this service will not promise
I do not promise rankings, citations, lead volume or a position inside any AI answer. No one controls what a language model says. Engines can misquote, misattribute or skip you, and their sourcing changes without notice. What I can do is strengthen the signals that correlate with being cited and show you, monthly, whether it is working.
Two more honest limits. First, schema is now about machine understanding rather than visible Google features: Google stopped showing FAQ rich results for all sites in May 2026, completing a phase-out that began in 2023, although FAQPage markup remains valid structured data (Google Search Central). Second, AI search optimization is rarely the first lever. If your organic base is thin, classic SEO and content come first; if you need pipeline this quarter, paid channels move faster. See the full picture on the services page.
Rank on AI for your industry
The prompts buyers type differ by industry, and so do the third-party sites engines cite in response. The questions a financial advisor’s prospects ask have little overlap with an M&A advisor’s. I work mainly with tax, M&A, consulting, real estate investing, capital raising and financial services firms, and the prompt set and mention plan are built per industry.
For a worked example of the vertical version, see how financial advisors can rank on ChatGPT. If your industry is not listed, the method still applies; the source map is what changes.
Who runs the engagement
Rank on AI is led by me, Christoph Olivier, founder of CO Consulting and a fractional CMO who has managed millions of dollars in ad spend and built a combined audience of over a million followers across social platforms. I publish original marketing data reports on this site, including the AI search readiness study, and this site’s own AEO cluster is built with the same method you would get.
In our work with growth-focused and professional-services clients, the pattern that keeps repeating is this: the brands already earning AI citations rarely got there by editing their own pages. They got there because they had spent years being mentioned, reviewed and quoted by the sources these engines trust, and our job was to make that existing authority legible to the machine through entity data, structure and honest schema. The clients who struggled were the ones who wanted the citation without the underlying credibility. We could not manufacture that, and we said so early.
If you want to hear how I think before booking a call, the video below is a talk on using AI on the sales side of a business. It is not a walkthrough of this service, but it shows how I approach AI as a practical growth tool rather than a buzzword.
Engine-specific AI search playbooks and research
Each major AI engine sources and cites answers differently. These free guides and studies cover the mechanics in depth, and the paid engagement applies them to your site. Start with the complete AEO guide if you want the full method, or go straight to the engine your buyers use most.
- Answer engine optimization: the complete guide (the full method)
- How to rank on ChatGPT (Bing index + earned media)
- How to get cited by Claude (Brave Search + primary data)
- How to rank on Perplexity (citation-first optimization)
- AI search readiness study 2026 (original research)
- AI search statistics 2026 (sourced data roundup)
- How to win at generative engine optimization
- AI search optimization across ChatGPT, Perplexity, Claude and Gemini
Frequently asked questions
These are the questions buyers ask most often before booking a call: whether this is only about ChatGPT, how long it takes, whether results can be guaranteed, how much it costs and how progress is measured. Each answer below is short on purpose. If yours is not here, bring it to the consultation and I will answer it directly.
What is the difference between AI search optimization, GEO and AEO?
In practice, very little. Generative engine optimization (GEO) and answer engine optimization (AEO) are two names for the same goal: being cited or recommended inside AI-generated answers. AI search optimization is the plain-English umbrella. All three build on classic SEO fundamentals, then add entity data, answer-first passages, original data and off-site mentions.
Is this just optimizing for ChatGPT?
No. ChatGPT is the largest assistant, but it retrieves through third-party providers such as Bing, Claude uses Brave Search, Perplexity runs its own index, and Gemini and Google AI Overviews use Google’s index. Winning on one does not carry you onto the others, so the engagement covers all of them from one baseline.
How long does it take to see results?
It depends on your starting point. Technical fixes such as unblocking a search crawler can show up once engines recrawl, while earned mentions and authority build over months. Established sites with real content usually move faster than new ones. I give you a realistic timeline on the scoping call after seeing your baseline, not before.
Can you guarantee my brand gets cited by ChatGPT or Google?
No, and you should be wary of anyone who does. AI platforms do not sell citation slots, and language models can misquote or skip sources. I can strengthen the signals that correlate with being cited, such as off-site mentions, entity consistency and clear answer-first pages, and report monthly on whether your share of voice is moving.
Do backlinks still matter for AI search?
Yes, but less than brand mentions. In Ahrefs’ 2025 study of 75,000 brands, branded web mentions correlated with Google AI Overview visibility at 0.664, compared with 0.218 for backlink count. Links still support crawling and authority, but being talked about on the sites engines already cite is the stronger lever.
How do you measure AI search results?
Three ways together: AI share of voice across a fixed prompt set, citation detail by page and engine, and AI referral traffic and conversions in a dedicated GA4 channel group. Some assistant traffic arrives without referrer data, so traffic alone undercounts the effect. The monthly report leads with share of voice for that reason.
How much does AI search optimization cost?
Pricing is scoped per engagement. The work depends on how much your site, entity data and off-site footprint already have, so a fixed package would either overcharge a strong site or underserve a weak one. Book a consultation and you will get a scope and a quote after the baseline review.
Should I block or allow AI crawlers?
If you want to be cited, allow the search crawlers: OAI-SearchBot for ChatGPT, Claude-SearchBot for Claude, PerplexityBot for Perplexity, plus Googlebot and Bingbot. Training crawlers such as GPTBot and ClaudeBot are separate; OpenAI and Anthropic both document that blocking them opts you out of training, not out of search citations.
Book a consultation
The fastest way to know whether Rank on AI fits is a short call. We look at your current AI visibility, your foundation and your buyers, and I tell you plainly whether this is the right lever now or whether your budget is better spent elsewhere first. There is no obligation and no guarantee pitch.