GREYWHALE / GEO
Make your business information clearer, better supported and accessible to AI search. We start with the questions buyers ask, the sources answers use and the facts your business can substantiate.
GEO stands for Generative Engine Optimization. It describes work intended to improve how a business is represented in AI-generated answers. Here, GEO does not mean geography, geotargeting or local SEO. An answer can mention a company without linking to it; a citation can link to a useful page without recommending that company. These are different outcomes.
A useful starting point is an evidence gap: an answer describes the wrong service, omits a relevant limitation, or cites an outdated source. The next action is to improve the underlying information and investigate where it is being retrieved. Repeating a keyword or asking a model to recommend a brand is not a substitute for this work.
Google says its existing SEO fundamentals apply to AI Overviews and AI Mode. Supporting pages must be indexed and eligible for a search snippet. No special AI schema or AI text file is required. Eligibility is not a promise of inclusion.
Google: AI features and your websiteOpenAI separates OAI-SearchBot, which supports search, from GPTBot, which concerns potential model training. These controls are independent. An access review should distinguish discovery from training preferences and check hosting rules as well as robots.txt.
OpenAI: crawler documentationThe following is a proposed project framework, not a claim of completed client results. Deliverables and responsibilities must be agreed before work starts.
Group real buyer questions into discovery, comparison and purchase decisions. For every observed answer, record the exact question, platform, date, language, market, cited URLs and whether the answer is accurate. Keep the original observation alongside the interpretation.
Map each information gap to an owned page or a relevant external source. Use approved specifications, implementation details, limitations and permissioned case evidence. Label examples as examples. Assign an owner and a review date to facts that can change.
Report brand mentions, linked citations and factual accuracy separately. State the number of questions and runs, platforms and markets tested, and any sampling changes. Track referral visits and qualified enquiries independently. A small prompt sample does not describe the whole market.
Record which pages changed, why they changed and what evidence would count as improvement. Keep an unchanged comparison set where practical. Treat before-and-after observations as directional evidence, because platform changes and other marketing activity can affect the same period.
Research note: Ahrefs studied 75,000 brands and found associations between external brand mentions and AI visibility. Its sample filters and correlation method matter: this does not establish that buying mentions, adding pages or copying a competitor will cause more citations. We use such research to form questions, not to promise a percentage uplift.
Ahrefs: study and methodologyFrom the discussion to the source: Jan-Willem Bobbink’s X post points to research on cooperative content optimization. The linked AutoGEO paper investigates experimental optimization methods; it is not evidence of GreyWhale client outcomes or a universal commercial playbook.
IS THIS THE RIGHT FIT?
For B2B companies, specialist service providers and brands whose customers ask complex questions before choosing a supplier. A useful starting point is a public website with real expertise and clearly defined services.
The final scope, deliverables, timeline and responsibilities are agreed in a project proposal. The items below describe the areas we can evaluate together.
WHAT THE WORK LOOKS LIKE
Review priority customer questions, existing pages and observable AI answers. Record the platform, date, location and cited sources so findings have context.
Clarify who you are, what you offer, who it is for and how services relate. Develop answer-led page outlines, comparison sections and a consistent company information set.
Check crawl access, canonical URLs, internal links and server-rendered content. Add structured data only where it accurately describes visible page content.
Maintain a repeatable question set. Review cited sources, referral visits and qualified enquiries alongside search performance; use the evidence to prioritise the next content update.
FROM BRIEF TO REVIEW
Identify the decisions your customers need to make and the evidence they expect.
Review your website and the sources currently answering those questions.
Publish clear answers, supporting evidence and useful links on the relevant pages.
Compare repeated observations and business outcomes, then adjust priorities.
PRACTICAL ANSWERS
No. Platforms independently select sources and generate answers. Our work improves the quality and accessibility of your information, but cannot guarantee citations, recommendations or rankings.
Yes. Crawlable pages, useful content and clear site structure support both organic search and AI discovery. GEO builds on those foundations rather than replacing them.
Your website, priority services, target markets, customer questions and approved evidence such as product documentation or case studies. Access needs are agreed after scope is defined.
YOUR NEXT STEP
Share your website, target market and business goal. We can use that context to discuss an appropriate starting point.
Talk about your project ↗