What Can a PR Agency Actually Do for GEO and AEO
By Jordan Baker, Founder of Fair Public Relations and former UK national journalist
More businesses are coming to PR agencies with a goal that barely existed a couple of years ago:
"I want my company to appear in ChatGPT."
It is a completely reasonable commercial objective.
If prospective customers increasingly ask AI systems which company to hire, which software to buy or which expert to trust, businesses understandably want to know how they become part of those answers.
The problem is that "AI visibility" can describe wildly different work.
One agency may mean technical SEO.
Another may mean content.
Another may mean schema and structured data.
Another may mean public relations.
Another may sell monitoring software.
Another may offer all of those things under one integrated service.
A business therefore needs to get beyond the label and ask a much more useful question:
What is this agency actually going to do?
A traditional PR agency can contribute to GEO and AEO by helping create clearer, stronger and more credible public information around a business, its people and its areas of expertise.
It can identify the positioning.
Develop media angles.
Prepare useful expert commentary.
Pitch relevant journalists.
Support interviews.
Secure earned editorial coverage when journalists independently decide the material is worth using.
What it cannot do is control ChatGPT, Google or another AI system.
It cannot guarantee citations.
It cannot guarantee recommendations.
And unless it genuinely offers those specialist capabilities, it should not pretend that ordinary media relations includes a technical SEO audit, schema engineering, AI crawler configuration or specialist generative-search analytics.
Why might public relations matter to AI search?
I come at this from journalism rather than computer science.
Before I ran Fair PR, I worked as a national journalist.
As a journalist, I did not quote somebody simply because their own website described them as a leading expert.
That claim was a starting point at best.
I wanted to know what they had actually done.
What experience did they have?
What could they explain?
Could they substantiate what they were saying?
Did their knowledge fit the story?
Were they useful?
That remains one of the clearest ways I can explain the role of PR in AI discovery.
A company website says what the company wants people to believe.
Earned media creates information outside the company's own channels.
That does not make every media article an endorsement.
A journalist quoting a person does not certify that they are the best professional in their field.
But somebody independent has made an editorial decision that the person's knowledge, experience, story or product was useful enough to include.
That creates another public source.
When AI systems increasingly retrieve and synthesise public information, I think it becomes commercially interesting that those independent sources exist.
Microsoft now describes grounding as a central part of AI search and has built reporting specifically around when publisher pages are cited in AI-generated answers.
That does not prove that one press article will cause one recommendation.
It explains why businesses are quite reasonably thinking about the wider public information surrounding them.
What would Fair do differently if the client's explicit goal were AI visibility?
I would ask more questions at the positioning stage.
What sorts of questions do you want to be a credible answer to?
When somebody is looking for a business like yours, what words are likely to describe the problem?
What attributes genuinely distinguish you?
Which locations matter?
Which sectors?
Which audiences?
What do you want the business to become known for?
And which of those claims can actually be substantiated?
Suppose somebody runs a nursery group and wants to become strongly associated with a particular positioning such as being a large independent nursery operator in the UK.
I would want to know whether that descriptor is factually defensible.
How many nurseries?
Independent in what sense?
Compared with whom?
Is there reliable evidence?
If it is accurate, then I would want the positioning to be consistently understandable across the wider campaign.
That might mean including the relevant context naturally in media biographies, press material or expert attribution when it genuinely helps explain who the spokesperson is.
It does not mean stuffing the phrase "number one nursery chain in the UK" into every quote because we hope an algorithm counts repetitions.
That would be poor PR and potentially misleading.
The principle is consistency, not keyword stuffing.
I would also care more about online coverage
If the explicit commercial goal includes AI discovery, digital publication matters.
Print coverage can be excellent PR.
Radio can be excellent PR.
Television can be excellent PR.
They may create considerable credibility, awareness and sales value.
But an online article creates a directly accessible public document.
That means, all else being equal, I would be especially interested in relevant online editorial opportunities for a client who specifically told me their objective included AI-search visibility.
I would still choose relevance over indiscriminate volume.
A specialist trade publication accurately describing the client as an expert in the precise field they want to own may be more useful than a completely unrelated mention on a more famous website.
The aim is not to collect URLs.
It is to build a coherent public profile.
What can a PR agency actually deliver?
A credible GEO-aware PR campaign begins with media positioning.
This means working out what the company and its people can realistically become known for.
If the founder wants to be recognised for artificial intelligence, do they genuinely have meaningful AI experience?
If a doctor wants to become a media source on health, which subjects fall inside their actual professional scope?
If an HR consultant wants to own workplace culture, what experience supports that?
Positioning matters because a scattered media profile can create lots of coverage without making the person easier to understand.
The second part is expert commentary.
Journalists constantly need knowledgeable people who can explain what is happening.
A PR agency can monitor relevant opportunities, identify developing stories, develop useful angles and help an expert communicate their knowledge clearly enough for publication.
The third part is proactive storytelling.
Not every useful media opportunity starts with a journalist asking a question.
A business may have original data.
An interesting customer story.
A founder experience.
A product or service that connects with something happening in the wider world.
A genuine business development.
Good PR identifies the editorial proposition rather than simply sending a company announcement.
Then there is thought leadership.
I use that term cautiously because it has become extremely overused.
To me, useful thought leadership is not manufacturing a controversial opinion for somebody because LinkedIn engagement is low.
It is identifying a real point of view grounded in experience and developing it into something another audience might genuinely learn from.
And finally, there is the core media-relations work: researching appropriate journalists, pitching them, handling responses, supporting interviews, supplying evidence and assets, following up appropriately and recording what actually gets published.
Those are tangible services.
Calling the same work "GEO PR" does not change what it is.
Work, interest and results are three different things
This distinction matters even more when AI visibility is involved.
A PR agency can carry out work.
It can write material.
Pitch journalists.
Develop ideas.
Research data.
Those are activities.
A journalist can show interest.
Request an interview.
Ask for images.
Say they are considering a story.
Those are opportunities in progress.
Then there is publication.
That is the actual earned-media result.
After that comes an entirely separate question:
Did an AI system later retrieve, cite, mention or use any of that information?
Those stages should not be collapsed together.
An agency should not say:
"We pitched your expert to twenty journalists, therefore we improved your ChatGPT ranking."
Nor should it say:
"We secured this article, therefore this later AI recommendation was definitely caused by us."
The evidence does not support either conclusion.
What might a real GEO-conscious PR campaign look like?
Take a hypothetical B2B consultancy specialising in recruitment technology.
The company has an experienced founder, but its online footprint consists almost entirely of its own website.
The founder wants prospective buyers researching recruitment technology in AI tools to understand that the company has genuine expertise in AI adoption, recruitment operations and hiring technology.
The first step would be positioning.
What has the founder actually done?
Which areas can they credibly discuss?
Which claims are supported by evidence?
What do buyers care about?
The agency might identify several coherent media subjects:
AI in recruitment
candidate screening
skills shortages
automation
hiring technology
employer adoption
the practical consequences for recruiters and candidates
Then we would look for editorial routes.
Perhaps the consultancy has anonymised data showing how employers are adopting automated screening.
Perhaps the founder can respond to news around AI hiring.
Perhaps there is an informed view on why a new technology is being misunderstood.
The agency develops the commentary and stories, approaches appropriate business, HR, recruitment and technology journalists and manages any resulting opportunities.
If articles are published, the business now has something it did not have before.
Independent public pages connect its founder with recruitment technology and AI.
That is a conventional PR result.
The business may then monitor whether its wider AI visibility changes.
But the PR agency should not invent the final step.
Unless there is evidence that a particular article was used by an AI system, we do not say it was.
Spectrum Search is the real-world version of the authority-building part
Spectrum Search is one of the clearest Fair PR examples of this sort of positioning.
The company could have spent its campaign continually promoting executive search.
Instead, we positioned Peter Wood around a wider but coherent set of subjects including recruitment, technology, AI and leadership.
He was exceptionally responsive and open-minded, which meant we could take advantage of a broad range of relevant opportunities.
The documented campaign generated 86 published media opportunities across 18 months.
That demonstrates earned-media authority building.
It does not demonstrate that those 86 articles caused a measurable improvement in ChatGPT recommendations.
We have never made that claim.
The distinction matters.
You can read the Spectrum Search case study for the media results themselves.
Can independent publications help brands become more visible to AI?
Potentially.
The safest way to answer this is to separate what we know from what we infer.
We know generative systems retrieve information from the web.
We know Bing now gives publishers direct data showing when pages are cited in supported AI experiences. Microsoft describes grounding queries as the phrases used when retrieving information subsequently referenced in generated answers.
We know Google's generative Search systems retrieve information using its Search infrastructure and that Google encourages original, expert-led, non-commodity content.
We know OpenAI says public websites can potentially appear in ChatGPT Search if they are accessible to its search crawler, while making clear placement is not guaranteed.
What we do not have is a universal formula saying:
one Daily Mail article = X increase in ChatGPT;
three trade articles = Y position;
ten backlinks = recommendation.
If somebody claims that formula exists, I would ask to see extremely good evidence.
Does the publication have to be enormous?
No.
This is one of the areas where conventional PR judgement and AI-search thinking fit together nicely.
Relevance matters.
Imagine a founder sells highly specialised software to architects.
A respected architecture industry publication may provide a much clearer external association between that founder and architectural technology than an incidental lifestyle mention on a vastly larger website.
I would never reject a major national opportunity simply because it was broad.
But I also would not chase a famous domain at any cost simply because somebody has decided its Domain Rating looks impressive.
Coverage should make sense.
Does a press mention need a backlink?
No.
I would prefer a good editorial backlink to the same article without one, all other things being equal.
A link can send referral traffic.
It can have conventional SEO value.
It makes it easier for readers to move from the publication to the company.
But an unlinked article still exists.
It can still name the company.
Identify the founder.
Describe their profession.
Associate them with a subject.
Quote their opinion.
Place all of that information on an independent public page.
We should not jump from that observation to claiming an unlinked mention is a proven AI-ranking factor.
But we should not pretend it is worthless either.
Our One Less Thing campaign shows the more conventional digital-PR side. It produced 38 published articles and 35 editorial backlinks in the first 12 months, alongside a documented increase in the site's third-party Moz Domain Authority metric from 0 to 19.
That is evidence of a stronger external link footprint.
It is not evidence that PR single-handedly caused every search change or AI result.
What is outside a traditional PR agency's remit?
This is where businesses need to ask difficult questions.
At Fair, Standard PR includes journalist-led strategy, angle development, content creation, media pitching, account management and campaign reporting.
It does not automatically include:
technical SEO audits
repairing indexing problems
website development
schema engineering
conversion-rate optimisation
managing Google Business Profiles
large-scale local citation management
search advertising
dedicated GEO monitoring software
complete customer-journey attribution
We undertake a lot of SEO and GEO work on our own website because it is our business.
That does not magically turn all those activities into components of the service a client buys from Fair.
I think agencies need to be unusually transparent about this because "AI visibility" makes it very easy to blur service boundaries.
When would I send a prospective client away to fix their website first?
I have done this.
If the website is unclear, that is a problem.
If important pages are not indexed, that is a problem.
If the site does not properly say what the business does, that is a problem.
If the owner has never connected Search Console or Analytics and has no understanding of what is currently happening, I would normally advise fixing those foundations.
If they have never done even basic work around the kinds of searches their customers make, I would want them to understand that too.
Sometimes a prospect needs PR.
Sometimes they need technical SEO.
Sometimes they need a new website.
Sometimes they need all three eventually, but in a sensible sequence.
I would much rather give somebody a list of things I think they should sort out before they spend money with us than take a retainer for a campaign built on broken foundations.
Google itself is very clear that GEO does not eliminate technical SEO
For Google's own generative Search products, a page has to meet ordinary technical requirements to be eligible to appear as a supporting link.
Google says there are no additional technical requirements specifically for AI Overviews or AI Mode, but the page still needs to be indexable and eligible within Search. It also says businesses do not need a special AI file or unique schema simply to appear.
That is why I would be sceptical of an agency selling a pile of technical "GEO hacks" while ignoring whether the website can actually be crawled.
What about llms.txt?
Fair currently uses llms-related website work as part of its own broader experimentation and information architecture.
But I would never sell a client an `llms.txt` file and imply it guarantees Google AI visibility.
Google's current guidance could hardly be clearer: Google Search does not use `llms.txt`, and having one neither helps nor hurts Google rankings or visibility.
Other services may use different mechanisms.
The wider ecosystem is evolving.
That is exactly why agencies should say which system a recommendation applies to rather than presenting one technical trick as a universal AI standard.
What should a GEO-focused PR campaign cost?
There is no credible universal "GEO PR" price because that label can describe completely different scopes.
A PR-only campaign may include strategy, media positioning, expert commentary, proactive stories, outreach and reporting.
Another provider may include technical SEO.
Another may add bespoke data research.
Another may provide specialist AI-monitoring tools.
Another may manage local listings, entity data and website development.
Obviously those services should not cost the same.
Fair's current Standard PR service is £797 per month in the UK, US$997 per month in the US and CAD$1,497 per month in Canada, with a three-month initial term followed by one full billing cycle's notice.
That is our PR retainer.
It is not a claim that £797 buys a complete technical GEO programme.
If a business needs technical search consulting alongside PR, I would tell them.
How should a business know whether the PR work is helping?
Start with the thing the PR agency actually controls.
Has it developed sensible positioning?
Is useful content being created?
Are appropriate journalists being approached?
Is there genuine journalist interest?
Is relevant coverage being published?
Are the subjects appearing in those articles aligned with what the business wants to become known for?
Those are legitimate PR measures.
Then separately monitor AI visibility.
Are relevant pages being cited?
Is the business mentioned?
Is it recommended for realistic buyer questions?
Is the description accurate?
Is AI-referral traffic arriving?
Are enquiries saying they found the business through ChatGPT or another assistant?
Those are different measures.
Blog 25 goes much deeper into precisely how I think businesses should track them.
What should you ask a PR agency before hiring it for GEO?
I would ask the agency to explain exactly what it is responsible for.
Is this PR?
Technical SEO?
Content SEO?
AI monitoring?
Or a genuinely integrated combination?
Ask what a normal month looks like.
Ask who does the work.
Ask how the campaign will be measured.
Ask what the agency can not guarantee.
Ask what happens if journalists do not like the original angle.
Ask whether technical website changes are actually included.
Ask what counts as a result.
And if they tell you they guarantee that ChatGPT will rank you first, ask how.
Transparency is more useful than futuristic terminology.
When do you need PR, SEO or both?
If your site cannot be indexed correctly, start with technical search/web work.
If your website is strong but almost nobody independent talks about your business or experts, PR may be the larger gap.
If you have excellent expertise but poor website positioning, you may need both.
If you are expecting a PR agency to guarantee ChatGPT recommendations, the first thing you need is a more realistic expectation.
There is no shame in using more than one specialist.
SEO and PR solve different parts of the visibility problem.
How Fair approaches GEO-related PR
We start with genuine media potential.
What does this business actually know?
What can its people credibly discuss?
What stories exist?
What evidence is available?
What does it want to become known for?
Then we shape those things around what journalists and their audiences actually need.
If a client's wider objective includes AI discovery, I would be more deliberate about the public associations we are building and especially interested in strong online editorial coverage.
I would want the company's description to be accurate.
I would want expertise consistently associated with the right subjects.
I would want important differentiators expressed naturally where relevant.
And I would want us measuring the wider digital picture rather than pretending the campaign ends when an article is published.
But I would still call the core service what it is: PR.
We do not need to rebadge ourselves as an artificial-intelligence consultancy to recognise that the environment PR operates in has changed.
Frequently asked questions
Can a PR agency improve ChatGPT visibility?
Potentially. PR can create credible external information about a company and its experts. However, an agency cannot guarantee that ChatGPT will cite or recommend a business.
Is GEO the same as digital PR?
No. Digital PR can contribute to a GEO strategy, particularly by creating online editorial references and links, but GEO may also involve technical search work, content, website architecture, data consistency and measurement.
Does Fair PR offer technical GEO or SEO services to clients?
Not as part of Standard PR. Fair's client service is primarily journalist-led earned media. We undertake significant SEO/GEO work for our own website, but that does not make technical optimisation part of every client retainer.
Can a PR agency guarantee AI citations?
No. Independent AI platforms control their own systems and outputs.
How many press articles are needed for GEO?
There is no evidence-based universal number. The relevance and clarity of the wider public information matters more than inventing a quota.
Do trade publications count?
Absolutely in PR terms, and potentially in AI discovery too. A specialist publication may create an especially clear association between a business and its field. That does not mean every trade article will necessarily be retrieved by an AI system.
Is an unlinked mention useful?
It can be. It still creates public third-party information about the business. A link adds referral and potential conventional SEO value, but an unlinked media mention is not automatically worthless.
How long should I assess the work?
Think in months rather than days. PR requires time to develop positioning, test angles and build media results. AI visibility should also be assessed as a pattern rather than through one screenshot.
Do I need PR and SEO?
Possibly. A technically weak website may require SEO/development work first, while a business with good search foundations but little independent authority may benefit strongly from PR.
Interested in building credible visibility through earned media?
Fair Public Relations helps founders, experts and businesses build a stronger independent editorial profile through Standard PR and Expert Commentary PR.
We develop positioning, stories and expert commentary, approach relevant journalists and manage the resulting media opportunities.
We do not guarantee AI rankings.
If your real problem is that very little credible information exists beyond your own website about what you know and what you do, that is where our work may be useful.
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 across journalism and PR. He writes about journalist relations, earned media, expert positioning, GEO, AI discovery and how businesses can build authority outside the channels they control.
View Jordan Baker's media expert profile
