TL;DR

For eight weeks, I’ve written about what look like eight unrelated problems—org charts, CFO conversations, HubSpot’s moat, a green owl on rented land. 

They were never eight problems. They were one: how to build trust in the age of AI. 

Trust isn’t a setting you switch on when you notice it’s missing; it’s a supply chain with stations you cannot skip, and the PESO Model® is the floor plan. 

The three shortcuts people try instead—automate it, fabricate it, disclose it—all failed in public last year. Here’s what the whole eight-week arc was actually building toward.

Key Insights

• Trust isn’t a setting you turn on. It’s what you actually do, the owned media that states it, the earned coverage that corroborates it, the shared media that repeats it, and the paid that ships it farther. 
• There are three shortcuts that almost everyone takes, including automation, fabrication, and disclosure. 
• Only 20% of people say they trust AI itself, and just 25% think they’d recognize AI-generated content if they saw it—so your audience has adopted the only rational defense available and assumes everything is synthetic until proven otherwise.
• Getting cited by an AI engine isn’t the finish line. Gartner found that 53% of consumers distrust AI search summaries, which means you face two audits: the machine’s and the skeptical human’s when they go to verify what the machine said. The good news is that one evidence base satisfies both.
• The 2026 Edelman Trust Barometer found business is now the only institution seen as both ethical and competent. Trust has drained out of everything else and pooled around companies, which makes spending it on a shortcut the most expensive decision available to you.

AI Visibility Starts With Trust

There’s a moment near the end of every talk I give that I’ve come to dread a little.

The Q&A is going fine. Someone raises a hand. And then comes the question, “Which tool are you using for AI visibility?”

I always grin and give the worst answer one can give, “It depends.”

I know what they want. They want a name. Ideally, a name with a free trial, a super-easy learning curve, and a dashboard. They want something they can expense before the end of the quarter, mention in a status meeting, and make them a hero within a day or two of using it.

And I have to tell them the truth, which is deeply unsatisfying, and it truly does depend. I use lots of AI tools for lots of different reasons and lots of different outcomes.

But the truth is deeper than that. Overnight success doesn’t come from my AI tool of choice. 

We’ve published content every week for nearly 20 years. That’s the tool we use for AI visibility. Not Claude or ChatGPT or Lovable or Grok. Those things help, for sure, but true AI visibility comes from publishing content every week for nearly 20 years. 

There is no easy button. No AI tool will do that for you. You can’t achieve overnight success; it’s not a thing. I’m sorry to be the one to tell you that, but that’s the truth. 

When I say this to an audience, I can visibly see the disappointment land. It’s not the answer anybody wants, and I understand why. It sounds like I’m being coy, or humble-bragging, or selling the slow thing because I don’t have the fast thing.

I’m not. I just don’t have a fast thing to sell you. I wish I did. I truly, truly do. But the fact is, nobody does. And anyone who says they do is lying.

Eight Weeks, One Problem

For the past eight weeks, we’ve talked about what look like completely unrelated topics. Au contraire, my friends! Au contraire. 

We’ve discussed:

• That CMOs don’t have an influence problem, they have an operating system problem. 
• A playbook for showing up inside the AI answer. 
• How HubSpot built a content moat, which we discovered when I ran them through a PESO Model® diagnostic lens. 
• Defending your budget in language the CFO uses. 
• What AI is already saying about your brand.
• The org chart that is quietly sabotaging everything you publish, 
• How Duolingo is building the internet’s favorite brand on rented land.

It seems sort of related because they’re mostly about AI, but they’re actually leading us to one big problem. 

Drumroll, please!

How to build trust in the age of AI. 

Trust is a Supply Chain, Not a Setting

We talk about trust as though it’s a state of being. 

You have it, or you don’t. Your brand is trusted, or it isn’t. 

And when it isn’t, we go looking for the thing that will turn it on—a campaign, a rebrand, a purpose platform, and now, an AI visibility tool.

Trust is not a setting. It’s not something you can turn back on after a stumble, or worse. 

Trust is a supply chain. 

It has stations, and the output of each one becomes the raw material for the next.

The first station is what you actually do. The product, the service, the way you treat people when it costs you something. If this station is broken, everything downstream is just well-distributed fraud, and no amount of marketing can rescue it. 

(We’ve watched clients try. It goes badly, and it should.)

The second station is owned media, where the claim gets made in a form that both humans and machines can find, read, and quote. 

This is where you say what you do, in public, on the record, over and over and over again.

The third station is earned media, where an independent third party confirms everything you’ve said about yourselves. 

Nobody has ever believed a claim just because a brand made it about itself. Not in 1970, not now, and certainly not in a language model built to check whether sources agree.

The fourth station is shared media, where actual humans repeat it in rooms you’re not standing in.

And the fifth is paid, which moves the finished product farther and faster, but cannot change what’s inside the box.

Paid. Earned. Shared. Owned. You’ve seen this floor plan before.

The reason the PESO Model® has outlasted a decade of frameworks is that the four media types map to the actual stations where credibility is built, in the actual order they must occur.

And you cannot skip a station. You can only decide whether to notice that you skipped it.

Three AI Visibility Shortcuts, and How Each One Fails

Because I always get the question about shortcuts or overnight success, and because we’re human beings who love instant gratification, I have three shortcuts people like to take that always go badly. 

Hopefully you can learn from one or all of them.

AI Visibility Shortcut One: Automate It

The first shortcut is to automate everything. And look, I’m a fan of automation. You can scale more quickly and do so much more if you automate. But not everything should be automated.

Just ask Klarna. 

By early 2024, their AI assistant was handling roughly two-thirds of customer service chats—the work of about 700 agents—and the company spent a year without hiring while its headcount fell by more than 20%.

The CEO at the time said they did this to cut costs, but they also sacrificed quality in the process. Because of it, Klarna started hiring humans again.

He said something worth remembering, “From a brand perspective, a company perspective, I just think it’s so critical that you are clear to your customer that there will always be a human if you want.”

I agree. You can’t build trust between a human and a machine. I’m not sure that’s possible. Not saying we won’t get there eventually, but we’re definitely not there now.

Automation made the interaction cheaper, but it did not make the brand more trusted. 

Efficiency and credibility are not the same currency, and one does not convert to the other.

AI Visibility Shortcut Two: Fabricate It

This second shortcut is worse, and it happened to a firm whose entire product is in being believed.

Deloitte Australia delivered a $300,000 report to the federal government—an assurance review of the system that automates welfare penalties, which is about as high-stakes as a document can get. 

Except…researchers found it contained a fabricated quote from a federal court judgment and references to academic papers that do not exist.

Oy. 

This can happen to any one of us. We’re all lazy and rushed at some point, and we start to trust the tools we’re using more than we should. 

Do not do that. AI hallucinates, and you have to check its work. Every time.

Deloitte ended up having to refund some of the very large payment they received for the report. 

Nobody set out to defraud anyone. Somebody used a tool that is extraordinarily good at producing the appearance of evidence—footnotes, quotes, citations, all the visual grammar of rigor—and skipped the station where a human checks whether the evidence is real.

AI Visibility Shortcut Three: Disclose it

And the last shortcut is one that is often debated in public.

The reasonable response to submitting a report to a client like that is to disclose, right? Label it. Tell people you used AI. Let it take the blame when something goes haywire. After all, honesty wins.

I wish.

Researchers at the Nuremberg Institute for Market Decisions ran a study and found that the disclosure penalty is real. 

When reviewing the same content side by side, people rate the AI-made version lower on credibility and emotional appeal, and they’re less willing to engage with the product. Even if it’s the same content and was actually human-made, but one version has an AI disclosure sentence on it.

In this case, transparency reveals a fundamental problem, but it doesn’t solve it.

Their companion research explains why the label lands so hard. Only 20% of people say they trust AI itself, and only 21% trust AI companies. About 44% know AI can produce marketing content, but just 25% think they’d recognize it if they saw it.

Therefore, your audience is suspicious, and they don’t trust their own ability to tell what is truly AI-created, so they assume it’s all synthetic until something proves otherwise.

Disclosure is necessary. I’m not arguing against it. But it’s a floor, not a strategy—and a brand that thinks a label buys back credibility has misunderstood what it lost.

Two Audits, Not One

Here’s the part I think most of the AI visibility conversation gets wrong, including in some of my own thinking.

We talk as though getting cited by an AI engine is the finish line. Show up in the answer, win the game.

Except the humans on the other side of that answer aren’t buying it either. Gartner found that 53% of consumers distrust or lack confidence in the reliability and impartiality of AI search summaries. 

Forty-one percent say AI overviews make searching more frustrating. A solid majority want a switch to turn the things off. 

(I personally find that strange because I far prefer AI overviews to regular search, but tomato, tomahto.)

So you’re not facing one audit. You’re facing two.

The machine audits you first, checking whether your claims are corroborated across sources it trusts. If you pass, you get named in the answer. 

And then the human, who is skeptical of the answer because it was AI-generated, goes looking to verify what the machine just told them.

The second audit determines whether you get the meeting, whether they click “add to cart,” whether they donate money, or whether they volunteer.

Which is oddly reassuring, when you think about it. 

It means the same evidence base serves both auditors. The earned coverage that makes a model confident is the same coverage that makes a skeptical buyer relax. The owned library a robot parses is the same library a human reads at 11 p.m. before they email you.

You are not building for the robots. You’re building the thing that has always worked, in a form the robots can also read.

Same strategy. New tools.

What Each Week Was Actually About

Now go back through our last seven weeks together, with the supply chain in mind.

1. The operating system article was about the factory itself—whether your function is built to produce anything coherent, or just to run channels in parallel and hope.
2. The org chart article was about what makes it off the line. Six approvers and a boring result are a production line that manufactures nothing worth inspecting.
3. HubSpot was the assembly station done right for more than a decade, until its owned library became a moat competitors can’t cross on a budget.
4. The visibility engineering playbook was the mechanics—how to structure what you make so it survives the trip.
5. The AI reputation article was the uncomfortable news that the audit has been running for a while, and that no one on your team was watching the results.
6. Duolingo was the cautionary tale about shipping enormous volume through a distribution channel you don’t own, with a thin record underneath. Famous and durable are different things.
7. And the CFO article was quality control—proving the line works, in the language of the person who funds it.

Seven stations. One system.

You Might Be Here If…

You might have a trust problem you’re trying to solve with tooling if…

• Your team can name three AI visibility platforms, but cannot name three earned media placements that corroborate your owned media from this quarter.
• The most recent article on your blog is a product announcement from five months ago.
• Your best proof points live in a sales deck, and nowhere a machine or a stranger could find them.
• Someone has proposed “getting into the AI answers” as a project, with a budget, a deadline, and no plan for what the answers would say about you.
• You’ve been asked to raise awareness, but the real problem is that the people who already know you don’t yet believe you.

That last one is the tell. Awareness problems and belief problems look identical on a dashboard and require completely different work.

The Unglamorous Part is the Whole Advantage

I’ll leave you with the finding that changed how I think about all of this.

The 2026 Edelman Trust Barometer reports that business is now the only institution seen as both ethical and competent—and that people trust their employer more than they trust government, media, or NGOs. 

Trust is draining out of nearly every other institution, and it has pooled around companies. Around us.

That’s not a trophy. That’s a liability with your name on it.

Because when you’re holding trust the rest of society has lost, spending it on a shortcut is the single most expensive decision you can make. 

The tools are seductive precisely because the real work is slow, boring, and impossible to put in a quarterly plan as a finished item.

Publish the thing. Pitch the story. Show up in the community. Measure what happened. Do it again next week when nobody claps. And again, when no one notices at all. 

That’s it. That’s the whole method, and it has an unfair advantage baked in. Your competitors can buy every tool you can buy, on the same afternoon, at the same price.

They cannot buy the years you spent being consistently worth citing.

So no, you can’t AI your way out of a trust problem.

But you can build your way out of one. You just have to start before you need it. The best time to plant a tree is 10 years ago. The second-best time is today. Start today.

Take Your Own PESO Model Diagnostic

If you want to know which station on your line is actually broken—and whether you have an awareness problem or a belief problem—take the PESO Model® Diagnostic. It’s free, it’s scored, and you can take it as often as you’d like.

And if you’re ready to build the whole system instead of patching one station at a time, the PESO Model® Certification walks you through it—owned authority, earned corroboration, integration, and the measurement that proves it to your CFO.

Or shoot us an email! We’d be happy to help you figure out where to start.

© 2026 Spin Sucks. All rights reserved. The PESO Model® is a registered trademark of Spin Sucks.