GEO for Small Businesses: Where Should You Start Before Spending Money?
By Jordan Baker, Founder of Fair Public Relations and former UK national journalist
If you run a small business, there is always another thing somebody tells you that you urgently need.
A website.
SEO.
Social media.
Email marketing.
Video.
Personal branding.
A podcast.
Then AI arrived, and now apparently you also need GEO, AEO, LLM optimisation and half a dozen pieces of software you had never heard of six months ago.
I actually think GEO matters.
We've invested very heavily in learning about it on Fair PR's own website.
I think AI search is already changing the way customers find businesses, and our own experience has made me more interested in shifting marketing effort towards GEO rather than treating traditional SEO as the only search channel that matters.
But that does not mean a small business should respond by immediately signing an expensive GEO retainer.
If you have limited time and money, there is a lot you can do before you pay anyone.
And most of it begins with something far less futuristic:
make your business easy to understand.
What does GEO actually mean for a small business?
At a practical level, it means thinking about how your business appears when people increasingly use AI systems as part of research and purchasing decisions.
Imagine somebody asks:
"Can you recommend an affordable interior designer in Manchester who specialises in small apartments?"
The system needs to work out which businesses exist.
Where they operate.
What they specialise in.
Whether there is evidence supporting that specialism.
Whether prices or positioning fit the request.
And which available public sources it should use.
That is why GEO cannot be reduced to one bit of code.
The full terminology is covered in our GEO vs AEO vs SEO guide.
For a small business, the much more useful question is:
What can I improve first with the time, information and budget I already have?
Step one: decide what you genuinely want to be found for
Before changing your website, answer this properly.
What do you sell?
Who is it for?
Where do you operate?
What are you genuinely good at?
What do customers normally hire you for?
What price range do you occupy?
Which types of client are a particularly good fit?
Which ones are not?
This sounds embarrassingly basic.
It is not.
AI recommendation systems deal with context.
Someone may not ask for "a PR agency".
They may ask for:
"a journalist-led PR agency for a UK health expert with a budget under £1,000."
That is completely different.
Fair PR is a good example
I do not want Fair recommended to everyone looking for public relations.
That would actually be evidence that the system had misunderstood us.
We are a relatively low-cost, founder-led, journalist-led boutique PR agency focused primarily on earned media.
We work particularly well with experts, founders, small businesses, professional services companies and certain B2B and consumer businesses.
If somebody needs international investor relations, huge corporate crisis infrastructure or a multimillion-pound integrated communications programme, I would prefer an AI system to recommend somebody else.
Good GEO should improve fit, not simply volume.
Write down your positioning in one paragraph
Try this:
We help [type of customer] with [specific problem/service] in [market/location]. We are particularly suitable for [use case] because [credible differentiator]. We are not designed for [important limitation].
If you cannot fill that in without ten marketing buzzwords, the website probably needs clarity before it needs advanced GEO.
Step two: look at your homepage like a stranger
Can somebody understand the business within a few seconds?
Not:
"We empower ambitious organisations to unleash tomorrow's possibilities."
What do you actually sell?
Where?
To whom?
If you are a plumber in Sheffield, say so.
If you are an online executive coach working with technology founders worldwide, say that.
If you are a Canadian accountant specialising in small ecommerce businesses, make that clear.
Clarity has become one of the most important lessons from Fair's own website work.
Check your important commercial pages
You do not necessarily need thousands of pages.
You do need useful ones.
Depending on the business, that might include:
your main services
important customer types
key sectors
locations
pricing
how the service works
case studies
credentials
detailed answers to common buying questions
These are sometimes loosely described as knowledge pages.
I would not get too hung up on the label.
The point is to create genuinely useful, specific resources around things customers actually need to understand.
Do not create pages simply because a keyword tool produced a list
Google's current guidance continues to prioritise people-first material that demonstrates real experience and depth, and specifically warns against producing large volumes of content primarily to manipulate search performance.
That is particularly relevant now AI can generate hundreds of superficially competent articles in minutes.
One excellent page based on genuine expertise may be much more useful than fifty interchangeable pieces saying what everybody else already says.
FAQs can be genuinely useful
I like FAQs.
Fair uses them extensively.
They force you to answer the actual questions buyers ask.
How much does it cost?
How long does it take?
Who is this for?
What is included?
What happens if it does not work?
Can you work in my sector?
Those are useful.
But don't add 100 questions simply because somebody has told you AI likes question marks.
Every answer should earn its place.
Step three: make sure search engines can actually find the website
This is where some small-business owners may need technical help.
Google currently says the same fundamental SEO requirements remain relevant to AI Overviews and AI Mode.
Important pages need to be indexable.
Crawling needs to be allowed.
Internal links should make content findable.
Important information should exist in text.
Google also recommends keeping relevant Business Profile and Merchant Center information current where applicable.
You do not need to become an engineer.
But you do need to know whether your site has serious problems.
Set up Google Search Console
If you have a website and have never looked at Search Console, I would start there.
Verify the site.
Submit or check the sitemap.
Look at what Google is indexing.
Look at impressions.
Look at the searches for which the site is already appearing.
Google explicitly recommends Search Console as part of establishing and monitoring an official website presence.
A sitemap does not force Google to index everything.
It gives the search engine information about the site.
Those are different things.
Set up analytics too
Google Analytics or another suitable analytics system helps you understand whether people are actually visiting the website and what they do afterwards.
Do not wait until six months into a marketing campaign and then realise there is no baseline.
Measurement does not need to be perfect.
It needs to exist.
What about ChatGPT crawling?
OpenAI currently says public websites can potentially appear in ChatGPT Search and advises website owners who want their content eligible to be discovered, cited and linked not to block OAI-SearchBot.
It also makes clear that placement is not guaranteed.
I would not recommend editing `robots.txt` blindly if you do not understand it.
A misplaced rule can block much more than you intended.
If you are unsure, ask somebody technical.
Do not ignore robots.txt, but understand what it is
`robots.txt` is an access-control instruction for crawlers.
It is not a GEO booster.
The goal is to make sure you are not accidentally preventing legitimate systems from accessing public content you want them to discover.
That is very different from adding magical AI text to the file.
What about schema?
Structured data can be useful.
Google explains that supported structured data can help it understand the meaning of page content and business information.
For example, organisation details, breadcrumbs and appropriate local-business information can support how Google interprets a website.
Use valid schema that accurately reflects the visible page.
Do not add fake credentials, reviews or information simply because markup makes them machine-readable.
What about llms.txt?
Fair has experimented with llms-related website work as part of our own evolving approach.
I would not tell a small business that `llms.txt` is one of the first things it must buy.
Google says its AI Search features do not require new machine-readable AI files or special AI schema.
The wider AI ecosystem may continue developing standards and conventions.
But if you have one day and no money, fixing an unclear services page is almost certainly more useful than obsessing over an experimental file.
Step four: get your public business information in order
Your website is not the only place customers or machines may encounter you.
Check relevant public profiles.
Business name.
Address.
Telephone number.
Website.
Opening hours.
Service description.
Professional credentials.
Founder name.
Social profiles.
Pricing where you choose to publish it.
They should not contradict each other.
Local businesses should pay particular attention to Google Business Profile
If you serve customers at a physical location or within a local area, claim and maintain your Google Business Profile.
Google says verified businesses can manage information such as address, contact details, category and photos across Google Search and Maps.
For a café, salon, tradesperson or local professional, this is foundational work.
What about directories?
Use relevant directories, not every directory on Earth.
A legitimate professional directory can be useful.
A respected trade association can be useful.
An appropriate local directory can be useful.
Yelp may make sense in markets where customers actually use it.
Industry-specific directories may be particularly relevant for professional services.
What I would not do is pay somebody to blast your details into hundreds of low-quality directory sites solely to create "AI signals".
Google explicitly lists low-quality directory and bookmark links among examples of potential link spam when used for ranking manipulation.
Relevance matters.
Reviews matter to customers even before we talk about GEO
Encourage genuine reviews on platforms your customers actually use.
Do not buy fake reviews.
Do not manufacture them.
Do not review yourself.
A real customer explaining a real experience is valuable because another human being considering the business can learn something from it.
Whether one particular AI system retrieves a particular review is a secondary question.
Start with customer usefulness.
Step five: publish useful information based on real expertise
This is where a small business can do a lot itself.
Think about what customers repeatedly ask.
Then answer it properly.
Not in two paragraphs written solely to rank.
Use what you genuinely know.
A recruitment consultant
Instead of another generic article called "Top Ten Interview Tips", perhaps explain:
why experienced candidates are rejecting certain roles;
where AI is genuinely changing recruitment;
how salary expectations differ in the specific market you work in;
or what employers misunderstand about a real skills shortage.
An interior designer
Explain a real trade-off.
Quartz versus natural stone.
When open shelving is a bad idea.
What actually adds cost to a kitchen renovation.
How to design a small room around awkward proportions.
Use your real experience.
A local restaurant
Customers may want very practical information.
Menus.
Dietary options.
Accessibility.
Parking.
Reservations.
Private dining.
Opening hours.
Children.
Dogs.
Special events.
High-quality photography.
Those things may be more useful than a generic weekly blog.
A professional adviser
Explain misunderstood questions inside your professional competence.
If the subject is regulated, keep the boundaries clear.
Authority does not come from pretending to know everything.
The best content often comes from what you already know
You do not always need external keyword research to begin.
Open your inbox.
Look at sales calls.
Look at customer support.
What do people keep asking?
What makes them hesitate?
What do they misunderstand?
What do you explain repeatedly?
Those questions are an excellent starting point for useful website material.
Step six: make your expertise verifiable
Do not just say:
"We are experts."
Show why.
Named author.
Professional background.
Relevant qualifications.
Years of experience where useful.
Case studies.
Real work.
Specific examples.
Previous media.
Customer evidence.
Useful original data.
This has always mattered to people.
The rise of AI gives businesses another reason to make it easy to understand.
Fair's biggest GEO lesson has been authority plus clarity
If I had to identify the things that have probably contributed most to Fair beginning to appear in ChatGPT and other AI-search experiences, I would put clear messaging very high on the list.
We became much more explicit about:
what Fair is
who we work with
what Standard PR costs
which countries we serve
which industries we know
what we actually deliver
where we are not the right agency
The second thing is authority evidence.
Case studies.
Real media coverage.
Detailed articles.
Expert profiles.
Evidence of actual PR work.
And we have consistently continued developing the website rather than treating it as a finished brochure.
Internal linking has also been a major part of the work
We have deliberately connected related content.
A sector page should not sit alone.
It can connect to relevant case studies.
Blogs can support commercial pages.
Expert profiles can reinforce specialist subject areas.
Pricing content can support buying decisions.
The site's important sections become connected rather than isolated.
Google specifically recommends making important content easy to discover through internal links.
That is basic SEO.
It also makes perfect sense for any retrieval system trying to understand a large website.
Do thousands of pages automatically help?
No.
Fair's website has grown substantially, particularly because our Media Coverage estate documents a large amount of earned editorial work.
I would not tell a small business:
"Fair has thousands of URLs, so you need thousands too."
That would be ridiculous.
Raw URL count is not the objective.
Useful information and proof are.
A tiny specialist consultancy may only need a comparatively small site to explain itself extremely well.
Step seven: decide whether you need external authority
At this point ask:
Does anybody independent talk about us?
If the answer is no, that does not automatically mean you need PR.
Plenty of successful businesses operate without media campaigns.
But if you want to build broader credibility around expertise, PR may become relevant.
## What can realistically become a media opportunity?
Useful expertise.
Original data.
A genuinely interesting founder story.
A noteworthy business development.
A distinctive product.
A customer or real-life story.
A useful service.
A strong opinion grounded in experience.
A local story with genuine community significance.
The mistake is thinking:
"I need GEO, therefore I need newspapers to mention my company."
Journalists do not publish articles because a business needs authority signals.
You still need something worth publishing.
Local PR can be perfectly worthwhile
A small local business does not automatically need The Times.
If you operate one café in Saskatoon, relevant local media may be far more commercially meaningful than a national British newspaper.
If you are a highly specialised B2B consultant, the right trade publication may matter more than general consumer coverage.
I would not claim either option is automatically superior for AI discovery.
Start with the audience and editorial relevance.
What can you reasonably do yourself?
| Activity | Usually possible internally? | When outside help may make sense |
|---|---|---|
| Clarify what the business does | Yes | When positioning is unusually complex |
| Improve main service information | Usually | Copy/content specialist for major rewrite |
| Correct public business details | Usually | Large multi-location footprint |
| Set up Search Console/analytics | Often | Technical problems or complex tracking |
| Build useful FAQs/resources | Yes | Specialist editing or regulated topics |
| Fix serious technical SEO | Depends | Developer/SEO specialist |
| Implement complex structured data | Depends | Technical specialist |
| Gather genuine reviews | Yes | No need to outsource authenticity |
| Pitch journalists | Possible | PR expertise/time/research may justify support |
| Track several AI systems at scale | Basic testing is possible | Specialist monitoring can save significant time |
The purpose of this table is not to convince you to outsource.
It is to stop you paying a specialist to do something you could reasonably sort out yourself in an afternoon.
If you have one day and virtually no budget, what would I do?
My instinctive order would be:
First, fix the website's clarity.
Make sure your main pages actually explain who you are, what you sell, where you work and who the service is for.
Second, expand useful knowledge.
Add the most important FAQs and create one genuinely useful buyer resource answering something your customers repeatedly ask.
Third, establish and check the foundations.
Search Console.
Analytics.
Business Profile if appropriate.
Sitemap.
Basic indexing.
Crawler access.
Fourth, clean up the public footprint.
Relevant directories.
Professional profiles.
Review platforms.
Inconsistent contact information.
Then test.
What does Google show?
What does ChatGPT say?
What looks wrong?
That is an enormous amount of progress for a business that started with nothing.
A realistic 30-day GEO starting plan
Week 1: understand the current picture
Write down the ten questions a real customer might ask before buying.
Search your own business name.
Search the founder.
Look at Google.
Look at relevant directories.
Ask a few sensible AI buyer questions.
Do not keep changing the wording until your company appears.
Record what actually happens.
Open Search Console.
Check Analytics.
Look at your main pages on a phone.
Identify factual errors.
At the end of week one you should understand the gaps, not necessarily have fixed them.
Week 2: fix what matters most
Improve the homepage.
Fix the most important service pages.
Correct pricing if it is wrong.
Add location information.
Improve contact details.
Update relevant external profiles.
Resolve obvious technical issues you understand.
Escalate anything complicated to somebody who does.
Do not spend the week adjusting tiny metadata details while the homepage still fails to explain what you sell.
Week 3: publish something genuinely useful
Pick one or two important customer questions.
Answer them properly.
Use real examples.
Add evidence.
Name the author.
Explain why they know the subject.
Link to appropriate supporting pages.
If one article takes the whole week and becomes genuinely useful, fine.
You do not need thirty.
Week 4: review authority and decide what comes next
Ask:
Are the technical basics okay?
Is the business now clearer?
Is there a substantial content gap?
Do we lack public proof?
Would customers benefit from reviews?
Is local visibility poor?
Do we have genuine media potential?
Then decide whether the next investment is:
nothing
more internal work
a developer
SEO
content
PR
integrated specialist
The purpose of the first month is diagnosis and foundation, not guaranteed ChatGPT rankings.
Should you ask ChatGPT directly what is wrong?
Yes, within reason.
This can be surprisingly informative.
Suppose you ask:
"Which affordable cleaning PR agencies operate in the UK?"
Your company does not appear.
Then ask:
"Would Fair Public Relations be suitable for a cleaning company?"
If the system responds:
"Yes, Fair PR works with cleaning businesses, has a relevant case study and fits that budget",
that tells you something.
The system may already understand the company but fail to retrieve it in the first shortlist.
That is different from the system saying:
"No, Fair does not appear to offer cleaning PR."
The second suggests an understanding/information problem.
The first looks more like a retrieval or competitive-selection problem.
We have encountered that distinction repeatedly in Fair's own testing.
Do not treat ChatGPT's explanation as an audit
There is an important caution.
Asking an AI system why it did not recommend you is useful exploratory research.
It is not a formal explanation of its hidden ranking algorithm.
The model may reason from the information available in the conversation.
Use the answer as a clue.
Then investigate.
How much should a small business spend on GEO?
There is no number I can responsibly give every business.
Your first investment may be nothing but time.
If the site is tiny and technically fine, perhaps the biggest opportunity is improving information.
If Search Console reveals serious problems, pay someone technical.
If your website is excellent but you have virtually no independent credibility, perhaps PR deserves investment.
If you already have both but need systematic multi-platform monitoring, a GEO specialist may add value.
Spend according to the problem.
What would I not spend money on first?
I would be extremely wary of:
guaranteed ChatGPT rankings
hundreds of generic AI-written pages
fake expert profiles
irrelevant directory packages
paid links disguised as editorial
"special AI code" that nobody can explain
dashboards with mysterious proprietary scores but no underlying data
large monthly contract before anybody has examined whether your website explains the business properly
Google specifically says no special AI markup or machine-readable AI file is required to appear in its generative Search features.
A provider may still have legitimate proprietary systems.
They should be able to explain what those systems actually do.
When should a small business hire a PR agency?
When there is enough material to work with.
You have credible expertise.
An interesting product.
A useful service.
A founder story.
Relevant data.
News.
A genuine customer story.
And you have time to participate.
PR is much harder when the founder will not answer questions, refuses to discuss anything except the product brochure or has no usable information on the website.
It may be worth fixing those things first.
Scenario one: local café
First priority: make sure the basics are absolutely correct.
Location.
Opening hours.
Menu.
Vegetarian/vegan options if relevant.
Accessibility.
Booking.
Photos.
Business Profile.
Reviews.
A customer asking an AI assistant "Where can I get vegetarian brunch near me?" needs practical information.
Possible PR opportunity: a genuine opening, unusual concept, founder story, notable menu, community initiative or event.
What can wait: complicated national GEO consulting if the fundamental local information is still wrong.
Scenario two: independent financial adviser
First priority: clarity and credibility.
Actual qualifications.
Authorisations.
Services.
Who they help.
Geography.
Professional boundaries.
Useful explanations of topics inside their competence.
Possible PR opportunity: responsible expert commentary around money, pensions, mortgages or financial planning where credentials fit.
What can wait: generic AI content claiming expertise on every area of finance.
Scenario three: recruitment consultancy
First priority: make the specialism clear.
Which industries?
Which roles?
Which markets?
Executive search or volume recruitment?
Candidate or employer focus?
Then publish genuinely useful insight based on what recruiters are actually seeing.
Possible PR opportunity: hiring data, salary insight, workplace trends, AI and recruitment, interviews, skills shortages.
What can wait: mass-generating generic "career tips" pages.
Scenario four: online product brand
First priority: useful product information.
Accurate name.
Price.
Availability.
Images.
Specifications.
Shipping.
Stockists.
Customer questions.
Merchant data where appropriate.
Possible PR opportunity: reviews, seasonal roundups, trends, founder stories, expert-backed content or genuine launch news.
What can wait: complicated authority-building campaigns if the product pages themselves are incomplete.
What Fair's own experience has taught me
The biggest lesson has not been a mysterious hack.
It has been clarity.
Fair's website increasingly gives systems and people concrete information rather than generic agency language.
The second has been authority.
Case studies.
Media coverage.
Detailed guides.
Expert profiles.
Actual evidence.
And third, I think consistent website development and internal linking have been extremely important.
We have not launched a homepage and left it sitting there.
We keep developing the information architecture and connecting genuinely related pages.
That process has coincided with rapid growth in Fair's visibility across conventional search and AI discovery.
I deliberately say "coincided with".
We are changing numerous things at once.
AI platforms are changing too.
It would be dishonest to pretend we can identify one button that caused the results.
GEO can move faster than traditional SEO, but do not expect instant results
In Fair's own early experience, some AI-search gains became visible remarkably quickly compared with the slow burn we often associate with conventional SEO.
That is one reason I think small businesses should take GEO seriously now.
But our experience is one business.
Your market may be different.
You may have much stronger competitors.
Your site may be older or newer.
Your industry may be more regulated.
The responsible promise is not:
"Do this for 30 days and ChatGPT will recommend you."
It is:
"Do this for 30 days and your business should be substantially clearer, more measurable and better prepared for whatever search and AI systems decide to do."
That is a worthwhile outcome by itself.
Frequently asked questions
What is GEO for small businesses?
It is the practical work involved in making a business discoverable, understandable and credibly represented as people increasingly use generative AI as part of search and buying decisions.
Can a small business do GEO without an agency?
Yes. Many foundational tasks, such as clarifying services, correcting business information, publishing useful answers and setting up basic measurement, can be done internally.
Is GEO expensive?
It does not have to start that way. Specialist technical work, content production or monitoring may cost money, but the first improvements may simply require time.
Does a business need a blog to appear in ChatGPT?
No. A blog can be useful for answering detailed customer questions, but it is not a universal requirement.
Can local businesses appear in AI recommendations?
Yes, potentially. Accurate local information, appropriate public profiles and genuine customer relevance matter, but recommendation is never guaranteed.
Do I need structured data or llms.txt?
Supported structured data can help search engines understand certain page information. Google says no special AI structured data or new machine-readable AI file is required for its AI Overviews or AI Mode.
Is Google Business Profile relevant?
For genuine local businesses, absolutely. It is a core way of managing how business details appear across Google Search and Maps.
Should I invest in SEO, GEO or PR first?
Diagnose the problem first. Serious technical issues suggest SEO/development. Weak owned information suggests content/website work. Strong foundations but little external authority may make PR more relevant.
How long does AI visibility take?
There is no universal timeframe. Some changes can be observed quickly while others develop over months. Independent platforms control inclusion and recommendations.
Can I guarantee ChatGPT recommends my business?
No.
What is the most useful first GEO task with no marketing budget?
Make sure the website clearly explains what you sell, who it is for and where you operate. Then check that public information is accurate and that basic search measurement is in place.
Ready to build visibility beyond your own website?
You may not need PR yet.
That is an important thing for a PR agency to say.
If your website is confusing or technically broken, fix it.
If the business information is inaccurate, correct it.
If you have never properly explained what you do, start there.
But once those foundations exist, you may discover that the next gap is external credibility.
That is where Fair Public Relations can help suitable small businesses turn real experience, expertise, services, products and stories into relevant journalist-led media opportunities.
Explore Small Business PR, Standard PR or get in touch if you want us to tell you whether we think there is actually a PR opportunity.
About the author
Jordan Baker is the founder of Fair Public Relations and a former UK national journalist with more than a decade of experience in journalism and PR. He writes about small-business PR, SEO and GEO strategy, AI search, earned media and practical ways businesses can build credible visibility without wasting money on work they do not yet need.
View Jordan Baker’s media expert profile
