We recently came across a story another agency shared that stopped us in our tracks. They had helped a brand through a product recall that was handled well, resolved fully, and followed by the natural fade of media attention. Google results were clean, customer concerns had been addressed, and from the outside, the crisis appeared to be firmly in the past.
Months later, however, the brand hit an unexpected wall. Sales were not rebounding as expected, and there was no obvious reason why. The press was no longer covering the issue, social media was quiet, and nothing new had surfaced.
The crisis ended but the damage did not
The problem turned out to be AI. When potential customers asked tools like Gemini or Claude about the company, the first thing mentioned was the recall, often described as if it were still happening. There was no indication that the issue had been resolved, no sense of timing, and no context to suggest the company had moved on. What should have been a closed chapter was being presented as a current risk.
This points to a reality many brands are only beginning to understand. AI does not move on the way people do. Reporters stop chasing a story, and search engines eventually prioritize newer information, but AI systems often draw from patterns in past data. If a crisis creates a strong signal and the recovery does not create an equally strong one, that moment can become the brand’s defining reference point.
You cannot correct AI you have to retrain the story
What makes this especially challenging is that AI cannot be corrected the way a news article can. There is no editor to call and no simple way to issue an update. You cannot ask a model to forget. Instead, you have to give it better information to work with.
In this case, the agency shared that progress came from clearly signaling that the recall was a resolved, historical event. By strengthening the brand’s digital footprint with accurate, authoritative updates and structured information, AI tools began to adjust how they talked about the company. Within about a week, the recall was no longer framed as breaking news but as a past issue that had been addressed.
AI has become part of the reputation landscape
The takeaway for communications and PR leaders is simple. Your reputation now lives in places you don’t fully control, and some of those audiences are not human.
As AI becomes a first stop for buyers doing research, crisis recovery cannot end when the headlines stop. Today, more than 2 billion people each month see AI-generated summaries in search experiences, making AI one of the most common first impressions a brand will ever have. For many decision-makers, those summaries shape perception before they ever visit a website, read a press release, or speak to a salesperson.
That means rebuilding trust today is not just about convincing customers, reporters, or partners. It also means making sure AI systems have the right information. If those systems are still focused on a past crisis and do not understand that it was resolved, brands can end up repeatedly paying for a problem they already handled responsibly.
This does not replace the basics of good PR like honesty, accountability, and clear communication. It builds on them. Modern crisis recovery includes clearly showing that an issue is closed by updating content, reinforcing timelines, and consistently sharing progress across the digital places AI pulls from. When that happens, AI tools are more likely to describe your brand accurately instead of getting stuck in the past.
Now that AI is often involved in how first impressions are formed, the real question is no longer whether your crisis is over, but whether AI understands that it is.
