From press coverage to AI visibility. What happens after the story is published?

There is a familiar point in most PR campaigns when the coverage report lands in the client’s inbox.

It contains links to articles, readership figures, reach, share of voice, and estimated impressions, all ways of demonstrating the scale and impact of a campaign. But there is now an obvious opportunity to go deeper and understand what happens to that coverage after it gets published.

We have recently been looking at what happens after that point. More precisely, we wanted to understand whether the body of content created as part of an earned media campaign could also contribute to how a company appears when people use ChatGPT, Claude, Gemini, Perplexity, Copilot, and other Large Language Models (LLMs) to research its market.

One programme we worked on with our partners Blue Train Marketing recently, confirms that it does.

The campaign

The client is a B2B financial services business operating in the banking and payments market. Mixology was brought in to implement an earned media strategy which would raise its profile and establish greater authority around the issues affecting its industry.

Blue Train Marketing, an agency specialising in digital marketing and search for clients in the fintech sector, was working on the company’s wider content and search presence. It also spearheaded a consumer behaviour study looking at attitudes towards retail banking.

For a PR agency, original research is valuable because it gives journalists something new to work with. We used the findings to create a media story and then developed a wider programme of editorials, bylined features, and opinion pieces aimed at the fintech, payments, and banking press.

The research campaign secured 11 pieces of media coverage over a seven-week reporting period. Thought leadership articles appeared in Retail Banker International and Finextra, alongside coverage in other international business and financial services publications.

This was recognisable PR: researching a subject, finding the news, developing the argument, and persuading independent publications that it was worth covering. There was no attempt to engineer articles for a chatbot, which made what happened to the client’s AI visibility while this work was underway all the more interesting.

The numbers moved

The business was being monitored across the UK, Germany, the Netherlands, and the US. Average AI visibility across those four markets stood at 5% at the beginning of the monitored period, and by August 2026 it had reached 14.75%.

Competitive rankings also changed sharply. The business moved from 78th  to 7th in the UK and from 107th  to 3rd in Germany. In the Netherlands it rose from 17th to 5th, while its US position moved from 20th to 7th. By August, the client ranked inside the Top 10 in every market being monitored.

Those are impressive numbers, but they do not prove PR alone caused the improvement, and that distinction matters. Over a longer period of time, Blue Train was developing the company’s search and content presence. More owned material was appearing online, and the website itself was becoming a stronger, more reliable source of information.

The research campaign was generating new content at the same time as Mixology was building the client’s profile through independent media, so several things were happening at once.

The useful evidence came when we looked at which sources were appearing behind relevant AI responses.

What the LLMs were finding

The source analysis was even more revealing than the rankings themselves – which showed AI responses drawing on the client’s product pages, research, editorial material, alongside articles and coverage published by third-party industry media.

That gives us a view of how the different parts of the programme were interacting. The company’s website was supplying information about its products and expertise; research and editorial content added depth, while independent publications created another layer of material about the business beyond the channels it controlled. AI platforms were finding material from across that mix.

This begins to explain why PR could matter much more to AI discovery than the industry originally assumed.

A company can optimise every page of its website, publish articles every week, and create detailed explanations of its products. All of this can make the business easier for search engines and LLMs to understand, but what it cannot do is independently verify its own reputation.

An article in a respected fintech publication exists because somebody outside the company decided the subject deserved editorial space. Research covered by the banking press has passed beyond the company’s own website, while an executive quoted as an expert in a news story has been selected by a journalist.

Those distinctions are familiar to anyone who has worked in media. They may now matter to machines as well.

Not all coverage will carry the same weight

There is a danger here for PR agencies.

Once earned media becomes associated with AI Visibility, the temptation will be to rebadge every form of coverage as an AI optimisation tactic. 

Press release syndication will become GEO, a backlink from an obscure website will suddenly be described as an LLM authority signal, and somebody will invent a proprietary score to prove its value.

We have been through versions of this before.

The more useful question is whether the coverage adds credible, relevant information around the subjects on which a company wants to be known. A meaningful interview in a specialist publication can do that, as can original research reported by respected media or an informed article which explores an important industry issue.

A copied press release appearing on dozens of low-value websites is likely to add very little.

This could force PR to become more disciplined about the difference between coverage volume and authority. For years, agencies have been able to produce reports showing that a campaign generated 100 pieces of coverage. The more difficult question has always been whether those 100 articles changed anything. AI Visibility provides another way of examining that question.

What happens after publication

Think about the cumulative effect of a good communications programme.

A company publishes a research report which receives media coverage, its CEO is interviewed about the findings, and the research prompts a bylined feature looking at the wider market. Six months later another journalist cites the original study in a different article.

The company begins to develop an identifiable body of authority around the subject, and those articles remain online, where search engines, LLMs, other journalists, and customers doing their own research can find them.

This is why the conventional coverage report increasingly feels like an incomplete description of what PR has achieved. Publication is an event, but the information created around it can have a much longer life.

For our financial services client, the evidence is that owned product content, research, editorial material, and independent coverage were all appearing among the sources supporting AI responses while its visibility improved.

We need a longer period of data before drawing broader conclusions. The models will continue to change, and different platforms do not behave identically, while AI Visibility is still an emerging area of measurement. The direction is hard to ignore.

A new responsibility for PR agencies

This has consequences for how communications programmes are planned. If a company wants to become associated with a particular area of expertise, PR teams should understand how the brand currently appears in AI-generated answers around that subject. They should know which competitors are being surfaced, and which publications are being cited.

That information can feed directly into a media strategy.

It might show that the company has plenty of owned content but virtually no independent authority, reveal that a competitor is dominating the publications which repeatedly appear as sources, or expose a subject the business believes it owns while LLMs barely associate the company with it at all. Those are useful communications insights.

They also move AI visibility away from being treated as another technical variation of SEO. Search and optimisation remain important because the brand has to be discoverable and its content intelligible, while earned media plays a different role by creating credible information outside the company’s own estate. The strongest programmes will understand both.

For our client, average AI visibility nearly trebled during the period measured, and the competitive ranking improved significantly across four markets. At the same time, the source analysis showed both owned and earned material helping support AI responses.

For a PR profession which has spent years asking how to demonstrate the lasting value of media coverage, that is worth investigating. The story may have a much longer life after publication than the coverage report ever showed.