Try a small experiment. Walk up to three people on your team and ask each one the same question: "What are you allowed to do with AI here?" Then compare the answers.
You'll get three different versions. One person quotes a half-remembered email. Another shrugs and says they use their judgment. The third didn't know there was a question. None of them point to the same place, because there is no same place to point to.
That's the quiet truth inside most organizations. The rules everyone assumes exist were never actually set.
The guessing problem
When a rule is clear, people follow it. When a rule is missing, people don't stop working. They guess, and they guess in the direction of getting the task done. The marketing lead pastes the campaign brief into a chatbot because it saves an hour and no one ever said not to. The developer drops in a chunk of code for the same reason. Each decision feels small and reasonable in the moment.
It feels reasonable because nobody handed them anything better. They're filling a vacuum the company left open. WalkMe's 2025 survey put a number on that vacuum: 51% of employees say they get conflicting guidance on when and how they're supposed to use AI at work. Half your people aren't ignoring the rules. They're trying to follow rules that contradict each other.
"Use good judgment" is not a decision
Faced with the gap, a lot of leaders settle on a comfortable non-answer: we trust our people to use good judgment. It sounds responsible. What it actually does is delegate the decision to whoever is busiest and least equipped to make it.
Judgment needs something to judge against. Without a shared line, "good judgment" just means everyone applies their own private standard, and those standards range from cautious to reckless depending on the person and the deadline. You don't have a policy. You have a few hundred individual policies, none of them written down, all of them quietly disagreeing.
Why the silence costs more than a bad rule
A clear rule that turns out to be slightly wrong is easy to fix. You spot the problem, adjust the line, move on. A missing rule is harder, because there's nothing to point at when something goes wrong, no shared expectation anyone violated. Everyone was improvising, so no one was accountable, so the same gap stays open for the next person.
And the gap isn't hypothetical. The same survey found 78% of employees already using AI tools the company never provided. The activity is happening at full volume right now. The only thing missing is the part where someone decided how.
The decision waiting to be made
This is not a technology question, and no tool will answer it for you. It's a plain organizational decision your company has never sat down and made: what is genuinely fine for our people to do with AI, what isn't, and who do they ask when they're unsure? Made properly, it comes out as a line anyone can follow: green for what's always fine, yellow for what needs a check first, red for what never goes in.
The AI Readiness Assessment shows you where that decision currently stands across the five dimensions that matter most, starting with whether anyone can actually answer the question you just asked your team. Five questions, about two minutes. It's built for organizations where nobody owns this yet, because that's who we build for: we put the governance system in place, and the person reading this is usually the one who ends up running it.
Your people are already deciding what's allowed, one rushed guess at a time. The real question is whether anyone with the authority to set the rule has decided first.
Keep reading
Part of a series on AI governance, the structure underneath the tools.
- Every AI Rollout Comes Down to Three Decisions. The worldview this whole problem sits inside.
- The 4 Stages of AI Governance Maturity. Where the guessing puts you on the ladder.