AI Strategy: Beyond the Hype, Toward Real Structural Change
In the last few months, I’ve found myself in more conversations about “AI strategy” than I can count. It’s become a catch-all phrase—one that can mean everything and, at the same time, almost nothing. The business landscape is full of pilots, proofs of concept, and tools that claim to automate, optimize, or disrupt. But I find myself returning to a much simpler, and perhaps more uncomfortable, question:If we were starting from scratch today, with AI at the core, what would our business actually look like?This is the lens I’ve brought to my work with fast-moving organizations. And it’s reshaped how I think about what it really means to “do AI” well.
1. The Real Work Begins with Leadership
Let’s be clear: AI isn’t an IT project. It’s a leadership and operating model challenge. The difference between incremental pilots and genuine transformation is the willingness—at the top—to rethink roles, workflows, and sometimes even the company’s identity. In my experience, the most successful efforts are those where leaders don’t just sponsor a project, but personally lean into the uncomfortable questions.
2. Choose Focus Over Fashion
There’s always a temptation to try everything at once. I see organizations running a dozen pilots, hoping that something will stick. But the real progress comes from making a handful of high-conviction bets—decisions that require a little courage and a lot of clarity. I often ask: If an AI-native competitor were starting here, what would they do differently? And more importantly: What’s stopping us from starting there ourselves?
3. Workflow Redesign, Not Just Enablement
Layering AI on top of legacy processes almost never delivers real breakthroughs. The substantial gains come when we’re willing to go back to first principles—redesigning workflows so that AI does the heavy lifting by default, and people orchestrate, judge, and steer. It’s less about adding another tool, and more about reimagining how work actually gets done.
4. Structural Moats Matter
Sustainable advantage in the age of AI isn’t about being the first to try a new tool. It’s about building assets that improve with use: proprietary data loops, reusable AI workflows, and domain-specific agents that become smarter over time. These are the structural moats that will separate the winners from the rest.
5. Culture Is the Ultimate Moat
In the end, no amount of technology can compensate for a culture that isn’t ready to change. AI fluency, a willingness to experiment responsibly, and visible, ongoing leadership commitment—these are what enable organizations to move from pilots to real reinvention. I’ve seen that the organizations that thrive are the ones where curiosity and trust run deep, and where learning is a team sport.
And yet, I don’t have all the answers. I still wrestle with the same questions many of you do:
How do we balance experimentation with governance?
How do we upskill teams for a world where AI isn’t just an add-on, but the default?
How do we avoid “AI theater” and focus on what actually moves the needle?
My ask for the community:
If you’re navigating these questions too, what’s working for you? Where are you stuck? Let’s compare notes—here or, if you prefer, in a confidential conversation.
Because in the end, AI strategy isn’t about having all the answers. It’s about asking the right—and sometimes uncomfortable—questions, and having the willingness to act on what you find.
If you’re wrestling with these challenges, I’d genuinely welcome your perspective.
Sabine Bankel, Chief of Staff as a Service, Founder 29 tasks