Generative Engine Optimisation
AI Search Optimization (GEO & AEO)
Search is splitting in two. Half your future buyers will never see a blue link — they will read an answer assembled by ChatGPT, Perplexity, Gemini or Google's AI Overviews, and act on whichever brands get cited. I engineer the entity signals, structured data, crawler access and content patterns that decide whether that citation is you or your competitor.
What's Included
Comprehensive solutions tailored to your specific needs
AI crawler access audit
Verified request testing for GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended and CCBot — robots.txt, WAF and edge rules.
Entity grounding programme
Organization and Person schema, sameAs consolidation, Wikidata and directory consistency across every public profile.
Citation-grade content architecture
Answer-first structure, question-shaped headings, tables, definitions and dated claims that models can extract cleanly.
Structured data engineering
Template-level JSON-LD across Article, FAQPage, Product, HowTo, LocalBusiness and BreadcrumbList, validated in CI.
Prompt panel monitoring
Scheduled testing across ChatGPT, Perplexity, Gemini and AI Overviews with citation share reported per prompt cluster.
AI referral analytics
Log and analytics segmentation so assisted sessions from AI engines are attributed instead of landing in direct traffic.
Problems I Solve
Symptoms this engagement is built to fix
Ranking on page one of Google while AI Overviews answer the same query using a competitor's content.
robots.txt silently blocking GPTBot, ClaudeBot, PerplexityBot and Google-Extended, making citation impossible.
No entity footprint: the brand and its founder exist nowhere in the Knowledge Graph or in LLM training and retrieval sets.
Content written as long narrative prose that models cannot extract a clean, quotable answer from.
Client-side rendered pages that AI crawlers — most of which do not execute JavaScript — see as empty.
Zero visibility into how often the brand is mentioned in AI answers, and for which prompts.
How AI answer engines actually choose what to cite
ChatGPT Search, Perplexity, Gemini, Claude and Google's AI Overviews do not rank pages the way classic search does. They retrieve a shortlist of candidate sources — usually from a live search index plus their own crawl — then generate an answer by extracting and synthesising passages, attaching citations to the passages that most cleanly support each claim. Three things decide whether you make that shortlist and survive the extraction step. First, retrievability: the crawler must be allowed in and must receive real HTML. Second, entity clarity: the model needs to know what your organisation is, what it does, and that independent sources agree. Third, extractability: the specific answer must exist as a short, self-contained, unambiguous passage rather than being implied across four paragraphs. Classic SEO gets you into the candidate pool. These three factors get you the citation.
Crawler access: the audit nobody runs
A surprising share of sites are invisible to AI search for a boring reason — their robots.txt or WAF blocks the crawlers. I audit access for GPTBot and OAI-SearchBot (OpenAI), ClaudeBot and anthropic-ai (Anthropic), PerplexityBot, Google-Extended (which controls Gemini and AI Overviews grounding separately from Googlebot), Applebot-Extended, CCBot (Common Crawl, which feeds many training sets) and Bytespider. Cloudflare's default bot-fighting rules and the newer AI-scraper toggles block several of these outright, as do many WAF managed rulesets, so robots.txt alone is not the whole picture — I verify with real requests using each user agent and check edge logs for 403s. Then we make a deliberate policy decision per crawler: publishers protecting licensable archives may want some blocked, but a services or SaaS business almost always wants all of them allowed, because a citation is free distribution.
Entity grounding: teaching models who you are
Language models reason about entities, not keywords. An entity is well-grounded when a consistent name, description and set of identifiers appear across your own site and across independent sources. On-site, that means Organization schema with logo, foundingDate, contact points and a complete sameAs block; Person schema for the founder and any named authors, with worksFor linking back to the Organization and sameAs pointing at LinkedIn, GitHub, Crunchbase, industry directories and social profiles; and a genuine About page written as a factual entity description rather than marketing copy. Off-site, it means the same name, address and description on every profile, plus mentions on sources models trust — Wikipedia and Wikidata where notability allows, industry associations, reputable directories, podcast and conference listings, and editorial coverage. Consistency matters more than volume: conflicting descriptions across sources make a model hedge, and a hedging model cites someone else.
Structured data for machine readers
Structured data has quietly become more important as SERP rich results have become less generous. AI engines parse JSON-LD to resolve entities, dates, authorship, prices and question–answer pairs without having to infer them from prose. I implement Organization and WebSite sitewide; Article with author, publisher, datePublished and dateModified on every editorial page; FAQPage wherever a real Q&A block exists — its SERP visibility fell in 2023 but LLMs still parse it heavily; HowTo for genuine procedures; Product with offers and aggregateRating for catalogues; LocalBusiness with geo coordinates for service-area businesses; and BreadcrumbList everywhere to expose site hierarchy. Every schema is emitted from the template layer so it cannot contradict the rendered page, and validated against Google's Rich Result Test in CI.
Writing content that models can quote
The single highest-leverage change most sites can make is structural. Answer the page's core question in the first 80–100 words in a complete, self-contained sentence that makes sense with no surrounding context. Use descriptive H2s phrased as the questions people actually ask. Define terms before using them. Put comparisons in real tables, specifications in real lists, and figures next to the entity they describe rather than three sentences away. Date and attribute claims — models strongly favour passages carrying an explicit date, a named source or a number. Add a short summary block near the top of long pages. Keep paragraphs to three or four sentences, because extraction works on passage boundaries. None of this makes the writing worse for humans; it is the same discipline good technical writing has always required, applied deliberately.
Measuring AI search visibility
You cannot manage what you cannot see, and Search Console does not report AI Overview citations separately. I set up a prompt panel — 40 to 150 buying-intent prompts covering your category, competitors, use cases and objections — and test them on a fixed schedule across ChatGPT, Perplexity, Gemini and Google AI Overviews, recording whether the brand is mentioned, whether it is cited with a link, which page was cited, and which competitors appear alongside. Server logs are filtered for AI crawler and referrer traffic (ChatGPT, Perplexity and Gemini all pass identifiable referrers on click-through) so real assisted sessions can be attributed in analytics. The output is a monthly share-of-voice number per prompt cluster, which turns GEO from a vibe into a metric you can hold work accountable to.
GEO and classic SEO are one programme, not two
Be sceptical of anyone selling AI search optimisation as a separate discipline with separate deliverables. Every engine studied so far draws its candidate set largely from conventional search indices, which means crawlability, indexation, site speed, internal linking and genuine topical authority remain the foundation. What GEO adds is a different emphasis on top of that foundation: entity work over keyword work, passage-level extractability over page-level relevance, off-site consensus over on-site optimisation, and citation share over ranking position. I run both as a single programme, because a site that gets cited by Perplexity but cannot be crawled by Googlebot is a site with two problems.
What You Receive
Deliverables & expected outcomes
- AI crawler accessibility report with per-bot verdicts and fixes
- Entity audit and sameAs consolidation plan
- Template-level JSON-LD implementation
- Rewrite brief for priority pages with answer-first structure
- Prompt panel of 40–150 buying-intent prompts with a baseline reading
- Monthly AI citation share-of-voice report
Illustrative Engagement
A representative scenario
The problem
A B2B SaaS ranked in the top three for its main category terms but appeared in no AI Overview and was cited by Perplexity for under 5 percent of its prompt panel, while a smaller competitor was cited constantly.
The approach
Cloudflare rules were blocking PerplexityBot and Google-Extended at the edge. We unblocked both, rebuilt Organization and Person schema with complete sameAs grounding, restructured 22 priority pages to answer-first format with comparison tables, and added FAQPage schema to the solution pages.
The outcome
Perplexity citation rate rose from 4 to 38 percent of the prompt panel in ten weeks, and AI-referred sessions became the fastest-growing acquisition channel over the following quarter.
Composite scenario based on typical engagements. Replace with a real, attributed case study once published.
My Process
A proven methodology for delivering exceptional results
Baseline
Prompt panel built and tested across four engines, plus a full crawler-access and entity audit.
Access & entities
Unblock crawlers, ship entity schema, and consolidate off-site profiles so every source agrees.
Content restructure
Priority pages rewritten to answer-first format with extractable passages, tables and dated claims.
Measure & iterate
Monthly prompt re-testing, citation share reporting, and content iteration against the prompts still lost.
Why Choose Me?
With 7+ years of experience and a track record of delivering results, I bring expertise, reliability, and a results-driven approach to every project.
- Visibility in the channel your competitors are not measuring yet
- Entity grounding that compounds across every engine at once
- Improvements that lift classic organic rankings simultaneously
- A real citation-share metric instead of anecdotes
- AI referral traffic properly attributed in analytics
- Senior consultant delivering directly — no junior handoff
Starting at
€397
Crawler access audit, entity grounding, structured data implementation, content restructuring brief and a monitored prompt panel with monthly citation reporting.
Frequently Asked Questions
Related Services
Further Reading
Related articles from the blog
Ready to Get Started?
Let's discuss your project and create a custom strategy for success.