The Permission Problem

A CFO I know spent $180K on AI tools for his company. Six months later, he's watching a 10% adoption rate and wondering what went wrong.

"We did the training," he said. "We hired consultants. We had town halls. What are we missing?"

After talking with him about the issue, I asked a few questions and told him what I thought the problem was. It wasn't about training or fear of technology. His team was paralyzed by something much simpler than technology: they didn't know what they were allowed to do.

The Invisible Boundary

Your team sees the new tools, attends the demos, and even completes the training modules. But they freeze when using the tools for real work, not because they can't figure out the interface or the system, but because they don't know their boundaries.

Can they upload client data to that new AI assistant?

Are they allowed to use it for customer communications?

What happens if the AI gives wrong information and they act on it?

Who's responsible when things go sideways?

These questions usually don’t get answered in training sessions. They get answered, or more often, don't get answered, in the messy space between policy and practice.

The Gray Zone Problem

Traditional software had clear boundaries. Your CRM system handles customer management. Your accounting software handled finances. The lines were obvious.

AI tools live in gray zones. They can write emails, analyze data, generate reports, draft contracts, create presentations, and answer complex questions. This flexibility is their strength and their adoption barrier.

When everything is possible, nothing feels permitted.

A marketing manager I spoke with recently had access to an AI writing tool for six months before she used it on anything important. Why? She wasn't sure if marketing copy generated by AI needed legal review, and she didn't know if her boss expected her to disclose AI assistance to clients. She couldn't find clear guidance on what level of editing was required.

So she kept using it for "safe" tasks: brainstorming and rough drafts that never saw daylight. The tool that could have transformed her team's output became an expensive notepad.

Permission Structures That Work

The companies succeeding with AI adoption don't necessarily have the best tools. They have the clearest permission structures.

Here's what they do differently:

They Define Acceptable Use Explicitly

Instead of generic "AI policies," they create specific use cases.

Here are a few examples:

• You can use the AI assistant for internal research and first-draft writing.

• You cannot use it for client-facing communications without human review.

• You can upload anonymized data for analysis.

• You cannot upload anything with personal identifiers.

They Make the Escalation Process Known

For every gray area, someone owns the decision. The marketing manager knows she can try AI-generated copy and run it past her director. The analyst knows he can experiment with AI data interpretation and check results with his team lead. Clear escalation paths eliminate decision paralysis.

They Make Failure Safe

This matters more than most leaders realize. Your team needs explicit permission to make mistakes while they're learning. Try it on low-stakes projects first. Document what works and what doesn't. Expect mistakes during the learning phase.

The Three-Layer Approach

The most effective permission structures I've seen use three layers:

• Green Zone: Tasks your team can use AI for without any approval. Internal research, brainstorming, first drafts, and data summaries for personal use.

• Yellow Zone: Tasks that require review or approval before implementation. Client communications, external reports, anything involving sensitive data, and decisions that affect other departments.

• Red Zone: Tasks that are off-limits until further notice. Anything involving customer personal data, legal documents, financial decisions, or confidential strategic information.

The key is making these zones explicit and accessible…not buried in a policy document but visible in the tools and/or within simple reference guides.

Creating permission structures isn't a one-time exercise. Your team will find use cases you didn't anticipate, the technology will evolve, and your comfort level will change.

The companies that succeed build learning into their permission systems. They regularly ask: "What new ways are people trying to use these tools? What permissions do they need? What boundaries do we need to adjust?"

They also recognize that permission isn't just about rules but about confidence. Your team needs to feel confident that experimenting within bounds is encouraged, not just tolerated.

The Leadership Question

If your AI tools aren't getting adopted, ask yourself: Does your team know exactly what they're allowed to do with them?

Not in theory. In practice. For the specific tasks they do every day.

If you can't answer that question clearly, neither can they.

And that's probably why your expensive tools are gathering digital dust.

P.S. I'm curious. When you rolled out your last major tool or system, what unexpected questions did your team ask? Those "gray area" questions usually reveal where your next permission gaps will be.

This is the kind of organizational challenge I help leadership teams work through regularly. If you're seeing low adoption rates on your AI investments, let's talk. Connect with me at ericbrown.com

If you found this post helpful, consider sharing it with another executive grappling with AI, technology, and data. If you want to explore AI and other Technology strategies, grab some time on my calendar, and let's chat.

Newsletter Recommendations