by Tyler Kelley
I ran an experiment last month. I asked five people at the same company to open whatever AI tool they normally use and type the same prompt: “Describe what we do, who we serve, and what makes us different. Write it as if you’re explaining it to a potential client for the first time.”
Five people. Five answers. Not one matched.
The descriptions weren’t wrong, exactly. They were each about 70% right. But they emphasized different things, used different language, positioned the company against different competitors. One mentioned an offering the company had quietly sunset six months ago. Another described a value proposition that sounded like it belonged to their biggest rival.
Nobody noticed.
This is what’s actually happening inside most companies right now, and almost nobody is talking about it. The AI adoption problem everyone’s trying to solve isn’t the real problem. The real problem is what your AI actually knows about you. For most companies, the answer is: nothing real.
A 2025 report from Netskope found that 72% of employees using generative AI at work are doing it through personal accounts. Personal ChatGPT logins, free-tier Gemini accounts, whatever they signed up for on their lunch break. The company has no visibility into what these tools are being told, what they’re producing, or what version of the brand is going out the door.
The security people call this “shadow AI,” and they worry about data leaking out. Fair enough. But there’s a different leak that’s harder to measure. Every one of those personal accounts is working from a different set of assumptions about who the company is. Sales trained their tool on old pitch decks. Marketing fed theirs a competitor analysis from last quarter. The founder’s assistant is working from a mission statement that was rewritten twice since then. Nobody coordinated. Nobody even thought to.
The output looks professional. That’s the trap.
AI-generated content is fluent enough that it passes a casual review.
Researchers from BetterUp Labs and Stanford gave this a name worth knowing: “workslop.” Content that looks polished but lacks the accuracy to actually move anything forward. 40% of U.S. employees had encountered it within the last month. Each instance cost roughly two hours of downstream cleanup by someone else. But those are the cases where someone noticed a factual error. When the problem is tone, or positioning, or strategy, nobody catches it. It just goes out the door.
That kind of incoherence is expensive, even if you never see the invoice.
The instinct, when this becomes visible, is to treat it as a training problem. Get everyone on the same AI tool. Run a workshop. Write a usage policy. IBM reported in 2025 that only 37% of organizations have any AI policies at all, so even that would put you ahead. But policies don’t solve this. A policy tells people what they can’t do. The actual gap is that nobody has given the tools anything real to work from. There’s no document, no reference, no shared source that every AI tool in the building can pull from when it needs to describe who you are and what you do.
The fix isn’t a policy and it isn’t a workshop. It’s a reference document that actually gets used. One version of who you are, what you do, how you talk about it, and what makes you different. Written once, updated when things change, and fed into every tool your team touches. Think of it less like a brand guide that sits in a shared drive and more like a cheat sheet you’d hand a sharp new employee on their first day. Except the new employee is every AI tool in your building.
Here’s how to find out if you need one. Send every person on your team that same prompt I used. Ask them to run it in whatever AI tool they normally use. Don’t tell them why. Then compare the answers. If they match, you’ve already solved a problem most companies haven’t identified. If they don’t, you just found the gap that everything else is built on.
Tyler Kelley is the Co-founder and Chief Strategist of SLAM Agency, helping organizations use AI to build visibility, strengthen relationships, and equip teams to deliver results that matter in an AI-driven future.