Hi Reader!

McKinsey's latest State of AI report shows 88% of organizations are now using AI regularly. But here's the catch: only 6% are actually seeing real value from it.

That massive gap tells us something important. Most teams haven't figured out how to work with AI in a way that actually helps rather than creates new problems.

The Tools Everyone Refuses to Give Up

I worked with an organization where teams were testing different AI tools and frameworks. Each team found something they liked and settled into using it. After several months of experimentation, leadership decided it was time to standardize—to pick what worked best for the whole organization.

That's when things got complicated. People had grown attached to their chosen tools. They'd invested time learning them, built workflows around them, and convinced themselves their approach was superior. When leadership suggested moving to a unified system, teams resisted. The resistance wasn't about the tools being better or worse—it was about loss.

What could have prevented this? Clear communication from the beginning. If you're letting teams experiment with AI tools right now, tell them upfront that experimentation has an endpoint. At some point, you'll need to evaluate together and choose what serves the whole organization. That clarity helps people stay open rather than territorial.

Why the Order of Work Matters

Research on how we use AI shows something interesting about our brains. When you ask AI to generate something first and then you edit it, your brain stays partially disengaged during the process. You're reacting rather than thinking.

But when you think through a problem first, form your own perspective, and then ask AI to add its input, your brain stays actively engaged. You're building on your thinking rather than accepting someone else's (or something else's) framework.

This matters more than it sounds. Teams that develop a practice of thinking before prompting maintain their critical thinking muscles. Teams that go straight to AI start losing that capacity over time.

What High-Performing Teams Do With AI

The organizations in McKinsey's 6% are handling AI differently in several ways:

They rebuild workflows completely. These teams aren't just plugging AI into existing processes. They're redesigning how work moves through their organization, which means AI becomes integrated rather than bolted on.

They set growth goals alongside efficiency goals. While most organizations focus purely on cost reduction, high performers also target innovation and growth. They're asking what becomes newly possible, not just what becomes faster.

They have clear validation rules. High-performing teams know exactly when AI outputs need human review and what that review should include. They've defined which decisions AI can handle, which need approval, and which require genuine human-AI collaboration.

Their leaders stay visibly involved. Senior leaders in these organizations use AI themselves and stay engaged when implementation gets messy. They don't delegate the transformation and disappear.

What Stays With Humans

Some responsibilities can't be handed off to AI, no matter how capable the technology becomes.

Final decision-making stays with humans. You can use AI to analyze options and surface insights, but the accountability for outcomes remains yours.

Empathy and human connection stay with humans. You can use AI to help with hiring processes, but you cannot use it to fire people or deliver difficult news. When someone is leaving your organization, they need a person who can see them as a person, not a system that processes them as data.

If you find these conversations uncomfortable, that discomfort is part of the job. Avoiding hard moments by inserting technology between yourself and your people doesn't make you efficient—it makes you absent when you're most needed.

Learn more from my conversation on the Developing The Leader Within podcast.

Three Things You Can Do This Week

1. Set the expectation now about tool choices

If your team is testing different AI tools, schedule a review date and tell them about it today. Make it clear that experimentation is welcome and temporary. You'll evaluate together and decide what works best for everyone.

2. Change the order of operations

Pay attention to whether your team asks AI first or thinks first. If they're going straight to AI for answers, shift the sequence. Have them work through their thinking, then use AI to build on it.

3. Define your human-only zones

Write down which decisions and interactions require human judgment and empathy. Share this with your team so everyone knows where the boundaries are.

Where We Actually Are

Three years into the generative AI era, most organizations are still figuring out the basics. The technology races ahead while we're trying to understand how it changes our work and our teams.

But the 6% who are capturing real value show us what works: redesigned workflows, clear validation processes, visible leadership, and a focus on what becomes possible rather than just what becomes automated.

Your job as a manager isn't to adopt every new AI tool or to resist all of them. Your job is to help your team maintain their thinking while using tools that can genuinely help, and to stay present for the moments that require human judgment and care.

Talk to you next Thursday

Daria

P.S. I talked about these ideas on the Developing The Leader Within Podcast check it out on Youtube