Trust · Strategy
The Language of Trust, Fifteen Years Later
When The Language of Trust was written, the premise felt urgent but arguable: audiences were becoming skeptics, and the language of business hadn't noticed. A financial crisis had just incinerated whatever benefit of the doubt remained, and companies were still communicating as if their word was their bond and everyone agreed.
Fifteen years on, nobody argues with the premise. They live it. So it seems like the right moment to ask an uncomfortable question about my own work: do the four principles still hold? Where the world has changed, how should the practice change with it?
The skeptics won
Let's start with the scoreboard. Public trust research keeps pointing to the same uncomfortable reality: belief is fragmented, audiences are more self-protective and institutional confidence is not something a communicator can assume. Edelman's 2026 Trust Barometer frames the environment as one of insularity and skepticism, not a simple left-right communications problem.
But something subtler happened alongside the decline. Skepticism stopped being a posture and became an operating system. The customer reading your product page has three other tabs open: a review aggregator, a Reddit thread, and a search for "[your brand] lawsuit." The employee reading the CEO's memo is simultaneously reading the anonymous comments about it. Verification isn't an occasional act of diligence anymore; it's ambient, instantaneous, and social.
For communicators, this changes the fundamental question. It's no longer "will they believe me?" It's "what will they find when they check?" — because they will check.
What changed since 2010
Three shifts matter most for anyone who writes or approves messages for a living.
First, the collapse of the official channel. In 2010, your press release still competed mainly with other press releases. Today it competes with a former employee's video, a customer's screenshot, and an influencer's hot take — all of which will be treated as more credible than you precisely because they aren't you. Your language now gets read against these sources, not instead of them.
Second, the arrival of AI-generated language. When fluent, polished, on-message copy can be produced in unlimited quantities for free, fluency and polish stop signaling effort or sincerity. Audiences are already developing an ear for the synthetic — the frictionless paragraph that says nothing checkable, commits to nothing, and could have been written about any company in any industry. Ironically, the corporate voice spent decades converging on exactly that style. AI just industrialized it.
Third, the speed of consequence. The gap between saying the wrong thing and paying for it used to be measured in news cycles. Now it's measured in minutes, and the payment is compounding: every misstep is archived, searchable, and attached to your next announcement.
Be personal, still
The first principle — be personal — asked communicators to talk about the audience's world rather than their own. Fifteen years later, this principle hasn't weakened; it has grown teeth. Generic language isn't just ignored now, it's read as evidence of automation, and automation is read as indifference.
The bar for "personal" has moved from tone to proof. It is not enough to sound warm or conversational. The message has to show it understands the audience's situation: the contract, the delay, the service failure, the health concern, the cost pressure or the real decision in front of them. Specificity is now a trust signal because generic language is cheap to produce.
Plainspoken in the age of AI
Be plainspoken was always the most popular principle in the room and the least practiced outside it. Everyone endorses plain language; then legal adds a clause, marketing adds a superlative, and the committee adds a paragraph.
What's new is that plainness has become a scarcity signal. When every company can generate unlimited smooth prose, the message that sounds like a person made choices — short sentences, a concrete number, an admitted trade-off — stands out the way a handwritten note stands out in a stack of printed mail. In dial sessions, the reaction to genuinely plain language has shifted from "refreshing" to something closer to relief. Audiences are exhausted, and plainness reads as respect for their exhaustion.
The practical test hasn't changed: read the sentence aloud to a neighbor. If you wouldn't say it across a fence, don't say it in a press release.
Positive and plausible
Be positive — framing what you do rather than dwelling in denial and rebuttal — survives intact, with one amendment. In an environment where your critics' version of events is a search result away, positivity without acknowledgment now reads as evasion. The sequence that tests best is acknowledgment first, then the affirmative case. Skip the first step and the second one is discounted entirely.
Be plausible remains, in my view, the principle that has aged best — and the one the AI era has made existential. Overclaiming was always expensive; every stretched superlative taxed the believable claims around it. Now the tax rate has gone up. Audiences armed with instant verification treat one implausible claim as a proxy for the credibility of the entire message, the entire brand, sometimes the entire industry. The strongest thing you can say is still the strongest thing your audience will believe. Not one word more.
Is there a fifth principle?
If I were writing the book today, I'd be tempted to add a fifth principle: be present. Trust used to be built in set pieces — the annual report, the campaign, the carefully staged announcement. It is now built, or lost, in the spaces between them: the reply to the angry customer, the answer given when the executive is off-script, the update posted when there's nothing new to report but silence would be read as hiding.
The organizations that earn the most durable confidence are usually the ones whose language is consistent across scripted and unscripted moments. The statement, the help-desk answer, the executive interview and the update when nothing has changed should all sound as if the same standard of honesty governs them.
Sources and limits
This article was refreshed on September 21, 2026 using public trust references including the 2026 Edelman Trust Barometer and Edelman's AI trust work. It does not add invented client research, quotations or proprietary test results; it applies the article's existing principles to a current communications environment.
FAQ
Does plain language mean oversimplifying?
No. Plain language means making the real trade-off legible. It can carry complexity better than jargon because it shows the communicator understands the issue well enough to choose.
How do you make a claim more plausible?
Narrow it until it can be supported. Replace universal promises with specific conditions, evidence, limits and next steps.
What to do on Monday
Principles are only useful if they change what ships. Three suggestions, drawn from what's testing well right now:
- Run the verification audit. Take your three most important claims and search them the way a skeptic would. If what they'd find contradicts or complicates the claim, rewrite the claim to survive the search — or address the complication yourself, first.
- Cut the synthetic sentences. Read your latest announcement and delete every sentence that could appear, unchanged, in a competitor's announcement. What remains is your actual message. It's usually one-third the length and twice as credible.
- Test before you trust. The gap between the conference room and the audience has never been wider. If a message matters, put it in front of real people before it ships. The dials still don't lie.
The skeptics won, and they aren't giving the ground back. But that's not a counsel of despair. It's a clarified assignment: in a world where nobody extends trust by default, the communicators who learn to earn it sentence by sentence hold an advantage that compounds. The language of trust was never really about words. It's about deciding to be believable — and letting the words follow.