A conversation with Shel Holtz in an episode of our For Immediate Release podcast in early June sent me down an interesting line of thought.
We were discussing the growing interest in making websites and online content easier for AI systems to understand. One example was The Economist, which has begun testing AI-ready versions of its web pages designed specifically for AI agents rather than human readers.
It's an intriguing development, but it also prompted a bigger question.
Are organisations beginning to acquire a second reputation?
Not the reputation people form through direct experience or what they read in the media, but the reputation AI systems assemble on their behalf before a human ever visits a website, speaks to a salesperson or picks up the phone.
The more I thought about it, the less convinced I became that this is really a second reputation at all.
AI doesn't create a reputation from scratch. It assembles one from the evidence an organisation leaves behind. News coverage. Customer reviews. Executive interviews. Research reports. Published content. Social conversations. Independent commentary. The consistency, or inconsistency, across all of it. That's an important distinction.
It's easy to look at developments like AI-friendly websites and conclude that organisations now need to optimise for AI. That's undoubtedly one way of interpreting what The Economist is doing with its AI-ready web pages.
It recognises that AI systems consume information differently from people and that organisations need to think about how those systems interpret their content. Certainly, making information easier for AI systems to interpret has value. Clear structure, authoritative sources and well-maintained content benefit everyone, human and machine alike.
But there's a danger in treating AI optimisation as though it's a new communication discipline.
Surely the goal isn't to optimise for AI. It's to optimise for credibility.
The evidence matters more than ever
If an organisation's public record is fragmented, contradictory or lacking substance, no amount of technical optimisation can improve the picture an AI system assembles. Increasingly, these systems synthesise information from multiple sources rather than simply repeat what an organisation says about itself.
That shifts the question communicators should be asking. Instead of wondering, "How do we look better to AI?", we should ask, "What picture does the totality of our evidence present?"
To me, that's the more interesting challenge. AI doesn't reward organisations for gaming the system. It rewards those that consistently leave behind credible evidence.
For years we've advised organisations that reputation can't be manufactured. It has to be earned through consistent behaviour, credible communication and actions that stand up to scrutiny.
AI hasn't changed that principle. If anything, it reinforces it.
Organisations don't have two reputations at all – they have one reputation viewed through two different lenses. One belongs to people. The other belongs to the systems that increasingly interpret organisations before people do.
If those two lenses produce very different pictures, the problem probably isn't the AI. It's the evidence.
AI may be changing how reputation is discovered. But it isn't changing how reputation is earned.
References:
- FIR 516: Your New Shadow Website (For Immediate Release, 1 June 2026)
- The Economist prepares for a two‑track internet: one for humans and one for AI agents (Digiday, 18 May 2026)
Photo at top by Laurenz Kleinheider on Unsplash.
