The reputation playbook the industry hasn’t updated

For twenty years, the online reputation management playbook rested on one assumption: content ages out. A negative article that ranked page one in 2015 would slide to page three by 2018, page six by 2020, and by 2023 nobody would find it unless they were actively digging. The whole industry — the suppression tactics, the microsites, the profile-stuffing, the "publish enough positive content and the old stuff disappears" model — was built on the physics of a ranking system that decayed over time.

That physics no longer holds. And the industry hasn't noticed hard enough.

There's a good piece in Search Engine Land today about a Midwest grocery chain whose ten-year-old customer service story got resurfaced by AI Overviews. The article had faded from Google's rankings years ago. The issue was resolved. The business had moved on. Then, seemingly overnight, the old story became a recurring source in AI-generated answers about the company. Not because it started ranking again. Because AI systems don't care about rankings — they care about whether a source is authoritative enough to cite.

That's a different game. And most of the reputation management industry is still playing the old one.

Suppression was a ranking strategy, not a truth strategy

The dirty secret of traditional ORM has always been that it never actually made the negative content go away. It just moved it far enough down the results that most people wouldn't find it. If you knew where to look, or you had a specific enough query, the old story was still there. But nobody scrolled to page four, so functionally it was gone.

The old model treated Google's ranking algorithm as the mediator between the reader and the story. The new model has removed that mediator.

AI Overviews don't scroll. They synthesise. When an AI system pulls together an answer about a company, it doesn't weight sources by their current SERP position — it weights them by domain authority, citation history, and the strength of the source as a reference. A ten-year-old article from a regional newspaper with real editorial standing beats a five-year-old profile page on a microsite you built to bury it. Every time.

The suppression tactics still technically work for their original purpose. You can still push results down the ten blue links. It just doesn't matter as much, because increasingly people aren't consulting the ten blue links. They're reading the AI summary and moving on.

The old model treated Google's ranking algorithm as the mediator between the reader and the story. The new model has removed that mediator.

Why AI systems keep going back to the old story

The mechanics here are worth being honest about, because they're not going away.

AI systems assemble answers by retrieving sources they consider trustworthy. Trustworthiness, for most of these systems, correlates strongly with things like domain authority, backlink profiles, editorial reputation, and how often a source has been referenced in the training corpus. A negative news article from a legitimate outlet ticks all those boxes. It was written by a real journalist, published on a real domain, indexed for years, and cited or referenced enough times to be part of the model's understanding of the topic.

Your response to that story — the microsite, the LinkedIn thought leadership post, the founder interview on a low-authority podcast — doesn't tick those boxes. Or at least doesn't tick them as convincingly as the original coverage.

So when an AI system is asked about the company, it goes back to the sources it trusts most. Which are often the sources that broke the bad news in the first place.

This isn't a bug. It's the system working exactly as designed. AI systems are supposed to prefer authoritative sources. The fact that the authoritative source is now out of date is not something the retrieval mechanism can easily know or account for.

The measurement problem, again

I keep coming back to this because it's the thing everyone underestimates. If a ten-year-old story is showing up in an AI Overview about your business, how do you know?

Your Google Search Console won't tell you. The article isn't ranking. Your brand mention tools will pick up new mentions but won't reliably surface the fact that an AI system is citing old ones. The current generation of AI visibility dashboards — as I've written about before — are mostly noisy at the citation level, and they're much better at telling you *whether* your brand appears than *what specifically* is being said about it or *which sources* are underpinning that description.

You basically find out because a customer tells you. Or because someone asks ChatGPT about your business and pastes the answer into an email. It's the same monitoring gap that runs through every AI search problem right now: the surfaces where the impact happens are the surfaces we can't reliably instrument.

Which means reputation problems that would previously have shown themselves in analytics (spike in searches for "[company name] scandal," drop in branded CTR) are now happening in a black box. You don't see the impact. You just start losing pitches, or getting weird questions in sales calls, or watching conversion rates dip without an obvious explanation.

What the industry hasn't updated

Here's what bothers me about the current ORM conversation. Most of the advice being given — including in the Search Engine Land piece — is still framed as an incremental patch on the old suppression model. Publish more citation-worthy content. Monitor AI platforms. Respond faster. Diversify sources.

The goal of reputation work has changed from ranking suppression to source authority displacement.

All of that is sensible. None of it is wrong. But it's still fundamentally the same playbook with "AI" bolted on.

The bigger shift, which almost nobody is naming clearly, is this: the goal of reputation work has changed from ranking suppression to source authority displacement. You're no longer trying to push a story down a list. You're trying to become a more authoritative source about your own business than the outlet that covered the negative story ten years ago.

That's a much harder problem. Because "more authoritative" doesn't mean "more recent" or "more optimised." It means genuine editorial weight — real journalism about you, in real publications, with real backlinks and real citations. It means your business becoming a topic that has enough legitimate coverage that the AI system has other places to go when asked about you.

This is not a content marketing problem. It is a PR and journalism problem. And most companies don't have a PR function that can operate at this level.

The awkward strategic implication

If source authority is the new suppression, then the businesses that will do well at ORM in the AI era are the businesses that were doing well at PR in the pre-AI era. Companies with genuine media relationships, a track record of being covered by trade press, a willingness to be quoted, and a habit of showing up in the industry conversation.

Companies that treated PR as a cost centre and reputation management as a tactical fix when things went wrong are going to find this new environment much harder to navigate. Because the tools they used to have — the microsite, the LinkedIn campaign, the SEO push — don't do the thing they used to do.

The other awkward implication: the negative story matters more now than it did when it was published. Ten years ago, a regional grocery chain getting a bad piece of local coverage was a temporary problem. It sat in the news cycle for a week, ranked for a while, then faded. Today, that same story is potentially a permanent reference point for how AI systems describe the business, unless and until enough authoritative counter-coverage exists to shift the source-weighting.

This has real implications for how businesses should think about press response in the moment a story breaks. The old calculus — "ride it out, it'll fade" — no longer applies. Because it won't fade. It'll get quietly cited in AI Overviews for years.

Where the honest limits are

I want to be careful not to overstate this. A few things I'd push back on if someone tried to use this piece to sell panic.

First, most businesses don't have an AI reputation problem. Most businesses have never had a significant negative story published about them at all. If that's you, this whole discussion is theoretical — your AI visibility issues are about being cited enough, not about being cited unfavourably.

Second, AI systems are still evolving fast. The retrieval logic that makes a ten-year-old story sticky today may be different in eighteen months. Google in particular has been actively working on recency signals in AI Overviews. The current asymmetry between old authoritative sources and newer counter-content may narrow.

Third, and importantly: no reputation strategy — old model or new — makes deserved negative coverage go away. If your business actually did the thing the story says it did, and did it repeatedly, no amount of PR sophistication is going to fix that. The best ORM strategy remains "don't be the kind of company that generates the story." Everything else is a rearguard action.

But for businesses in the middle — companies with one or two legacy stories that misrepresent where the business is now, or that made a real mistake once and have genuinely moved on — the strategic response has to change. Suppression is a wasting asset. Source authority is the compounding one.

What this actually means for how you work

If you're running a business with any exposure to legacy negative coverage, the practical shift is straightforward, even if it's not easy.

Stop investing in tactics designed to move old content down the rankings. That money is going toward a problem that matters less than it used to. Redirect it toward getting genuine, substantive, source-quality coverage of your business in outlets that AI systems will treat as authoritative. Trade press, industry publications, credible independent commentary — the kind of thing you can't buy but you can earn if the underlying business is doing interesting work.

And accept that ORM is now on a longer timeline. Not "publish six positive articles and wait for the algorithm to notice" — but "spend two to three years becoming a company that legitimate publications cover, so that the corpus of authoritative sources about your business is dominated by present-day reality rather than a decade-old incident."

That's a hard sell to a CFO who's used to ORM being a project with a defined end date. It's the accurate one anyway.

The industry will get around to updating the playbook eventually. Most of the practitioners currently selling ORM haven't yet, because the old model still sort of works and the new model requires capabilities most agencies don't have. But the businesses that get ahead of this — that stop thinking about reputation as a ranking problem and start thinking about it as a source authority problem — are going to be dealing with a much smaller mess in three years than the ones who don't.

The old stories aren't fading. Plan accordingly.