Authenticity Isn't a Tone of Voice Anymore — It's Evidence

For a few years, "be authentic" meant sounding like a person. Drop the corporate language, write like you talk, show some personality. It worked because it was hard enough to do that most brands didn't. Then writing like a warm, relatable person became something anyone could generate in four seconds, and the signal stopped signaling.

The trust numbers are moving the wrong way

Recent 2026 research points in the same direction. A Gartner survey of 1,006 US consumers, run in May and June, found 35% now rely on influencers less when shopping because of AI. Optimizely's UK research reports roughly six in ten consumers saying AI has made brand marketing less trustworthy. Clutch found 53% of consumers say they're less likely to buy from brands they know use AI in their social content.

These are stated attitudes, not purchase behavior, and they should be read that way. But the pattern across them is hard to ignore: people aren't rejecting AI as a tool. They're discounting content that feels mass-produced, because it's become cheap to produce and easy to suspect.

Tone was never the proof

Here's the real shift. Authenticity used to be judged by how something sounded. When sounding human is free, that stops working as evidence. What's left to judge by is what a reader can actually verify: whether the claims are specific enough to be wrong, whether the customers are real and nameable, whether the people behind the work are visible, whether the brand behaves the same way on a Tuesday as it does in a launch campaign.

That's a higher bar, and a fairer one. It also quietly punishes the shortcut we described in AI isn't a brand strategy: content that's polished, fluent, and interchangeable, with nothing underneath that couldn't have been said by the competitor.

Specifics are the new warmth

The brands holding trust tend to do a few unglamorous things. They name real results instead of "trusted by industry leaders." They let customers say specific, slightly inconvenient true things, the same dynamic that makes reviews raw material for how a business gets described. They show the process and the people, not just the finished surface. And they keep saying the same true things for years, which is exactly what a campaign-by-campaign voice can't fake.

None of this requires abandoning AI. It requires knowing which parts of the work can be accelerated and which have to start from something real: a position, a customer conversation, a result you can point to.

The test

Take your last ten pieces of brand content and cross out every sentence a competitor could have published unchanged. What's left is your evidence. If it's thin, the fix isn't a better tone of voice. It's having more to show.

This is an Express post — the second of our four stages: Define, Express, Build, Grow. If your content sounds right but has never named a real customer, a real result, or a real person, Express is where that gets fixed.