AI Is Rewriting Company Culture Before Policies Catch Up

By Amara Osei

At workplaces that have implemented AI, 24% of employees say their culture improved over the past year — while 25% say it got worse. That near-even split comes from Gallup's 2026 research, and it makes one thing clear: AI alone doesn't decide the outcome. The choices leaders make about how AI gets used do.

The core problem, as the analysis puts it, is that "employees are finding ways to incorporate AI into their work faster than many organizations can create policies, norms or expectations around its use." An employee asks AI a question they once would have discussed with a colleague. A presentation that used to take all day is ready before lunch. The habits take hold before anyone has decided whether they serve the company's stated values.

Is your AI culture forming before your AI policy?

AI used to research, draft, analyze, summarize and solve problems is reshaping how people work together. As new habits settle in, expectations about speed, collaboration and quality shift. The leadership question, the analysis argues, is no longer "Should our people use AI?" It's: "Are the ways people are using AI strengthening the culture we say we want or unintentionally working against it?"

Deloitte has a name for what happens when organizations obsess over how people use AI but ignore its effect on how they work together: "cultural debt." It builds up quietly, and leaders prevent it by translating existing company values into concrete AI-era behavior.

What tradeoffs are leaders actually making?

The analysis identifies three tradeoffs that signal to employees what an organization really values:

• Speed vs. critical thinking. A report that once took five hours now takes one. Someone still needs to question the findings and decide what they mean. If the answer is "produce more reports," little time remains for thinking — and output quality suffers.
• Access to information vs. collaboration. An employee can ask AI a product question instead of calling R&D. That saves time, but they lose the conversation. R&D never hears what customers are asking, and cross-team problem-solving gets harder.
• Efficiency vs. workload. Time saved by AI can fund harder problems, new skills or cross-team work. Or it can simply mean more work is expected — and burnout risk rises.

"AI doesn't make those choices. Leaders do," the analysis notes — and over time, those choices become the unwritten rules of work.

How do you turn values on the wall into daily behavior?

The framework maps five values to specific AI practices:

• Collaboration: Let AI handle straightforward questions; make time for conversation when a colleague's experience would improve a decision.
• Critical thinking: Treat an AI answer as a starting point — check facts, ask what's missing, decide what the information means.
• Learning: Give people room to experiment and discuss openly what worked and what didn't.
• Trust: Make clear when people must disclose AI use, who checks the work and who owns the final result.
• Strategic thinking: When AI saves time, protect space for harder questions before adding more tasks.

Two field examples show the approach in action. One manager had a team use AI to prep for a client meeting, then brought everyone together beforehand to stress-test it: What did the tool assume about the client? What did someone on the team know that wasn't in the summary? The team walked in better prepared — and reinforced the habit of learning from one another before trusting an AI answer.

In another case, a team used AI to sketch options for a new project, then gathered colleagues from other departments before choosing a direction. Those colleagues caught practical problems the team missed and explained how the change would affect their work. AI got things moving; the shared decision made it better.

What should leaders do with the capacity AI returns?

The analysis closes with a reframe: the question isn't only how much work AI can take off people's plates, but what leaders want people to do with the time and attention it gives back. Bringing AI into an organization, it argues, deserves the same attention as hiring a new employee or standing up a new department — think through how it fits and how it affects the people already there. That care, the argument runs, is how companies build culture by choice rather than by default.