For most of the last decade, earned media was the budget line nobody fought for, squeezed whenever a paid campaign needed more spent. Gartner’s last forecast flips that story: by 2027, it predicts PR and earned media budgets will double.
The driver is where people now go to decide what to believe about a brand. As generative AI tools replace traditional search as the first stop for discovery, more than 95% of the links cited in AI-generated answers point to unpaid content, while paid media barely registers. Reputation is no longer something bought during a campaign window: it’s built over years through content already in the world. Gartner’s predictions describe an infrastructure some companies already have, and others don’t.
1. Owned content becomes the only currency of visibility.
Buried inside Gartner’s headline number is a detail most comms teams have been ignoring for years: press releases are the least cited source in AI-generated answers. Indeed, more than 95% of the links cited by AI engines point to unpaid content, with 27% coming directly from earned media.
These models are built to weigh credibility the same way a sceptical reader would. A press release is a company describing itself: nobody has checked it, corroborated it, or repeated it independently. A trade publication running the same claim, or three unrelated outlets converging on the same fact, reads entirely differently to a model trained to look for consensus across independent sources.
That’s also why Reddit and Wikipedia now punch so far above their weight in AI answers. Neither looks like marketing. Wikipedia gives models a stable, cross-checked reference point; Reddit gives them something closer to live testimony, what people are actually saying, unfiltered by a comms department. Both read as consensus, which is the exact signal these systems are built to chase.
This means that most of what brands have already published was never built to be cited, with no hard numbers, no outside validation, and a phrasing too vague for a model to lift cleanly. That’s the real reason budgets are meant to double: not more production, but a different form of it.
It’s also the starting point of any serious audit. Before writing a single new page, the real question is which of a brand’s existing pages, statements or articles are already showing up inside AI answers and which have simply gone invisible and why. The Editorialist’s expertise in Editorial Strategy gives companies a clear picture of what is lacking in their GEO governance.
2.Reputation has to be continuously monitored, not measured after the fact.
Gartner’s second call is blunter than the first: narrative intelligence is moving from a nice-to-have to a necessity because disinformation is accelerating faster than most comms functions can respond to it.
A distorted narrative that used to spread through traditional press had natural friction built in through editorial cycles, fact-checking, and a news day that eventually ended. A narrative distorted through an AI engine has none of that. Once a false or misleading claim gets picked up and repeated across enough sources, it starts showing up as the answer itself, not as one version among several. Brunswick has been making a similar point about corporate reputation in its recent work on AI risk: the old rule of ‘respond within 24 hours to keep control of the narrative’ was built for a media cycle that no longer exists. By the time a narrative has hardened inside an AI answer, the correction window has usually already closed.
That’s the gap between narrative intelligence and ordinary media monitoring. Monitoring tells you what was said about you last month. Narrative intelligence is meant to catch the weak signal before it turns into a topic anyone has to manage.
Gartner breaks the discipline into three pillars: detection, contextualisation, and rapid response. Detection means picking up a signal while it’s still small. Contextualisation means knowing that signal is actually worth it, if this is a fringe complaint or the early shape of something that scales. Rapid response means having a position ready before the story is fully formed.
The Editorialist calls this Editorial Intelligence, and it works the same way. We help you with structured briefings, delivered on a set cadence, flagging emerging trends, regulatory movement or shifts in perception before they go mainstream: the same three pillars, built into a rhythm a comms team can actually act on rather than react to.
3.Communications has to prove its value with the same rigour as any other business function.
The third lesson we can take out of Gartner’s predictions is about a problem AI has simply made impossible to keep ignoring: comms reporting is shifting from retrospective summaries to predictive analytics, because the old way of justifying the budget has stopped working.
The real issue sits one level up from any dashboard. Finance and the executive committee have treated communications as a cost centre for years because comms has rarely handed them numbers that plug into the same models as sales pipeline, retention, or CAC. While everyone else at the table walks in with a forecast, comms walks in with a recap, describing activity, not impact.
This is exactly the gap The Editorialist’s custom reporting is meant to close: metrics built to function as impact indicators, not activity logs. The Editorialist’s AI Visibility Score or Editorial Authority Score tools are a quantification of something boards now care about directly: how present a brand is inside the systems that increasingly mediate its reputation and how credible it is. That’s the language finance understands: a number that moves, that can be forecast, and that ties back to a business outcome rather than a media recap.
4.Content structuring becomes infrastructure, both internally and externally
The fourth rupture in Gartner’s report looks, at first glance, like a separate story from the first: by 2028, 75% of employees will use chatbots rather than intranets to find internal information, while personalisation by profile replaces the one-size-fits-all memo.
However, an internal chatbot is only as good as what it’s reading. Ask it a policy question and it will pull from whatever’s sitting in the knowledge base, whether it be outdated procedures, half-updated pages, or three conflicting versions of the same document, if that’s what’s there. Employees will trust the answer regardless because it comes wrapped in the same confident, conversational tone whether the underlying content is current or three reorganisations out of date.
Organising a knowledge base demands the same semantic-structuring discipline as external AI visibility: clear, sourced, current, easy for a model to extract cleanly. Structuring content so ChatGPT cites you correctly and structuring content so your own internal assistant answers a benefit question correctly are the same skill.
The personalisation piece adds a layer most teams skip past too quickly. Gartner’s segmentation by role, geography, and priority is a research problem first. You can’t tailor a message to ‘engineers in APAC’ or ‘frontline managers three months into a reorg’ until you actually know what those groups care about, where they get their information, and what’s currently getting lost in translation between them and leadership. Segmentation without that groundwork is just guessing with better labels.
That’s exactly the gap the semantic/GEO audit section of The Editorialist’s Audience Insights offer is built to close. The same audit that maps external content gaps is run against the internal knowledge base, then paired with the audience research needed to actually know who you’re segmenting for before you try to personalise anything.
Conclusion
The four ruptures Gartner’s report displays share one root cause. Budgets double because unstructured, self-declared content doesn’t get cited. Reputation risk compounds because nobody catches the weak signal before it hardens. Finance treats comms as a cost centre because the metrics never speak its language. Internal chatbots misfire because the content underneath was never built to be read by a machine, let alone trusted by one.
The fix, every time, is the same: content that’s structured enough to be extracted, governed enough to stay accurate, measured enough to prove impact. The Editorialist helps you close a gap that already exists today between companies that built this foundation years ago and those still hoping the old way of producing content holds up.
