What generative engine optimization is
Generative engine optimization (GEO) is the work of making a business easy for generative AI tools to reach, understand and name when someone asks for a recommendation. The tools are ChatGPT, Google AI Overviews and AI Mode, Perplexity, Claude and Microsoft Copilot. They write one answer instead of listing ten links, often with links to the sources they used.
I am based in Seekonk, Massachusetts, and I do this work for businesses and organizations across the United States and Canada, on whatever platform the site runs on: Wix Studio, WordPress, Shopify, Squarespace, Weebly or any other builder.
How GEO differs from SEO and from AEO
Local SEO gets your pages and Business Profile ranked. Answer engine optimization (AEO) shapes a passage on your page so it can be quoted as the direct answer. GEO asks three other questions: can each AI tool reach your site, is it certain who you are, and do sources other than your own website confirm it.
Google says that optimizing for its AI features is still SEO and needs nothing special. The other tools run their own crawlers under their own rules. That is where most GEO-specific work sits.
AI crawler access: who is visiting and why
Each vendor documents its own user agents, and they do different jobs. Some gather training data. Some build a search index. Some fetch a single page because a person asked a question. Blocking the wrong one can remove you from an AI tool's search results without you knowing. This list comes from each vendor's own documentation.
OpenAI: OAI-SearchBot. Surfaces websites in ChatGPT search. OpenAI says sites that opt out will not be shown in ChatGPT search answers, and that it is not used for model training.
OpenAI: GPTBot. Crawls content that may be used to train OpenAI's models. Disallowing it signals that your content should not be used for training.
OpenAI: ChatGPT-User. Visits a page when a person's request in ChatGPT calls for it. Because a person started the action, OpenAI notes robots.txt rules may not apply.
OpenAI: OAI-AdsBot. Checks landing pages for ads submitted to ChatGPT. It only matters if you advertise there.
Anthropic: ClaudeBot. Collects content that could contribute to model training. Blocking it signals that future material should be left out of training data.
Anthropic: Claude-SearchBot. Crawls to improve search results in Claude. Anthropic says disabling it may reduce your visibility in user search results.
Anthropic: Claude-User. Fetches pages when a person asks Claude a question. Disabling it stops Claude from retrieving your content for that person.
Perplexity: PerplexityBot. Surfaces and links websites in Perplexity results. Perplexity says it is not used to train foundation models.
Perplexity: Perplexity-User. Fetches a page at a person's request and, Perplexity says, generally ignores robots.txt.
Google: Googlebot. The same crawler that feeds Google Search feeds AI Overviews and AI Mode. A page must be indexed and eligible to show a snippet.
Google: Google-Extended. A robots.txt token, not a separate crawler. It controls whether your content is used for Gemini training and grounding. Google says it does not affect inclusion or ranking in Search.
Three practical points follow. First, you can stay visible in AI search and still opt out of model training, because those are different user agents. That choice is yours, and I set the rules the way you decide. Second, robots.txt is only one gate. OpenAI and Anthropic publish the IP addresses their bots use, and a firewall, security plugin or CDN that blocks those addresses will keep the bots out even when robots.txt allows them. Third, changes are not instant. OpenAI says a robots.txt update can take about 24 hours to register.
What I do, step by step
Access audit. I check robots.txt, robots meta tags and firewall or CDN rules against each user agent above.
Entity clarity. One exact set of business facts, stated identically everywhere your business is described.
Structured data. Organization or LocalBusiness markup in JSON-LD, with sameAs links to your real profiles, matching the visible page.
Quotable content. Plain statements of what you do, where, for whom and with what credentials, kept as text. I also add first-hand detail only your business can supply, which Google's guidance calls non-commodity content.
Third-party mentions and reviews. I identify the legitimate places your business should be described and help you get listed correctly.
Mention checks. I run a fixed set of prompts in each tool and record what comes back, before and after the work.
Entity clarity, with a worked example
An AI tool can only recommend a business it is sure exists as one thing. Take a heating and cooling contractor as an example. The website footer uses the legal name ending in LLC. The Google Business Profile adds the words and Plumbing. An old directory listing shows a phone number from before the company moved. The Facebook page still has the previous address. The website never says which towns the company serves.
A person can work out that these are the same company. A system reading thousands of pages has to guess, and it may treat them as two businesses, repeat the old phone number, or leave the company out because it cannot confirm the service area.
The fix is unglamorous:
Choose one name, one phone number, one address or service area and one set of hours as the record
State them in plain sentences on the site, on a contact or about page, as text
List the services and the towns covered, by name
Correct every listing and profile to match the record exactly
Add Organization or LocalBusiness structured data that repeats the same facts, with sameAs links pointing to the real profiles
Google's documentation says some Organization properties are used to tell one organization apart from another. Other vendors do not publish how they resolve this, so I do not claim more than that. Agreement across sources is simply the condition under which any reader, human or machine, can be confident.
Where third-party mentions come from for a small business
A small business does not need press coverage. It needs to be described accurately in the places that already describe businesses like it:
Your Google Business Profile and the reviews on it. Google's guidance says Business Profiles can help your services be visible in AI responses as well as regular results.
Chamber of commerce and local business association member directories
Trade association listings and manufacturer or supplier dealer locators
State or municipal license lookups, where your trade is licensed
The franchise brand's own location finder, if you are a franchisee
Local news coverage, event sponsorships and community organizations you really support
Review platforms your customers already use for your trade
No vendor publishes which of these sources it weighs. Google does say that chasing inauthentic mentions is not as helpful as it might seem, and I do not buy mentions or reviews. The aim is that everything true about your business can be found, and that it agrees.
GEO for multi-location and franchise brands
Multi-location businesses and franchise brands have the entity problem many times over. The brand is one entity. Each location is another, with its own address, phone number, hours and service area. AI tools get this wrong when location facts live only in a store finder script, or when a franchisee's own site disagrees with the corporate listing.
Every location gets its own page with its facts as text, not only a pin on a map
Location names follow one pattern everywhere, on the site, in Business Profiles and in directories
The brand's location finder and each franchisee's listings are checked against each other
Crawler access is decided once, at the brand level, so one location's settings do not hide the rest
I have built more than 30,000 location pages, the Christmas Decor franchise site, and I have worked with more than 50 lighting companies. The page system is covered under programmatic SEO and location pages.
A repeatable way to check whether you are being named
You can do this yourself, and I encourage it.
Write a fixed list of prompts a customer would really type. Include a recommendation prompt (who installs holiday lights near a named town), a comparison prompt, and a direct prompt about your business by name.
Run each prompt in a new conversation in ChatGPT, Google AI Mode, Perplexity, Claude and Copilot.
Record the date, the tool, whether you were named, what was said about you, whether it was correct, and which sources were cited.
Where a fact is wrong, follow the cited source and correct it there.
Repeat the same prompts after changes are made and compare the two records.
The same question can return a different answer the next time it is asked, so this is a sample and not a ranking report. Two other checks support it. Server logs show whether the user agents listed above are actually fetching your pages. Google Search Console reports traffic from its own AI features, and Google's guidance warns that outside tools do not have access to its internal numbers.
What is known and what is not
No one can guarantee placement in an AI answer. OpenAI's own help page says placement is not guaranteed. No vendor publishes a formula for which businesses get named.
Google says structured data is not required for its AI features. I add it because it states your facts without ambiguity and supports regular rich results. I do not claim it gets you into ChatGPT.
llms.txt is a proposal, not an adopted standard, and Google says you do not need AI text files to appear in its AI features. I do not sell it as a fix.
What a model learned in past training cannot be edited. Changes show up first in tools that search the live web.
Why the work is still worthwhile
Every step here is also sound practice for regular search and for customers. A site that crawlers can reach, facts that agree everywhere and real reviews help you whether the person finds you through Google, a map listing or a chatbot. Nothing in this work depends on a trick that could stop working.
What I need from you
Editing access to the website and to robots.txt, plus whoever manages your hosting, firewall or CDN
Access to your Google Business Profile
The correct record: legal name, the name you trade under, address or service area, phone, hours and services
A list of the profiles, directories and memberships you know about
Your decision on AI training crawlers, once I have explained the options
What GEO will not do
Guarantee that any AI tool names or recommends you
Make up for missing reviews, a weak reputation or a service area you do not really cover
Control what an AI tool says. It can still state something wrong, which is why the checks are repeated.
Replace local SEO. Both rest on the same crawlable pages and accurate business facts.
Who it is for
Service businesses. Contractors, home services and lighting companies that people now ask AI tools to recommend by town.
Healthcare practices. Accurate facts about providers, services and locations matter most here.
Franchises and multi-location businesses. Many locations, one brand, and business facts that have to agree for every one.
Nonprofits. Clear statements of mission, who is served and where, so referrals reach the right people.
Questions to ask any GEO provider
Which AI user agents will you allow or block, and what does each vendor's documentation say it does
What do you guarantee, and how, given that OpenAI itself says placement is not guaranteed
Where will mentions come from, and are any of them paid for
How will you show me whether anything changed
How Meta Monster fits in
Meta Monster is a Chrome extension I built, a live SEO and AI page auditor. I use it to check AI crawler access and answer readiness on a page before and after changes. It is free to install from the Chrome Web Store, so you can run the same check yourself.
Ask for a free proposal
Call or text 508-244-0403, email joshua@websitedesignma.com, or use the contact page. I reply within 24 hours and the proposal is free. If a deadline is real, such as a lighting company heading into the holiday season or a business with an opening date, ask about an expedited timeline.