Executives keep asking the same question. What’s our AI strategy? Brian Evergreen has a blunt answer. There isn’t one. There is only strategy. AI fits inside it or it doesn’t.
The author, advisor and former Microsoft U.S. AI strategy lead has spent years watching companies chase tools before they know where they’re headed. His message lands harder now than ever. Organizations pour resources into pilots and use cases. Results disappoint. Inertia wins.
Vision first, then everything else
Evergreen made the case clearly in a May 21, 2026 conversation with GeekWire’s Todd Bishop. “Use cases are the friend of engineering, but the enemy of strategy,” he said. “Instead of being AI first, you need to be value first.” (GeekWire)
That distinction matters. Companies typically start with a problem. They identify a use case. They grab low-hanging fruit. Progress stalls. The pattern repeats across industries. Without a vivid picture of the desired future, decisions stay tactical. Budgets go to incremental gains. Real transformation slips away.
Vision supplies the force. “Vision is the only force with enough momentum to overcome organizational inertia,” Evergreen told the audience. No vision. No genuine strategy. No ability to make choices that compound over time. Short sentences. Long consequences.
His thinking appears consistently. In a 2023 CIO article he wrote that ChatGPT functions as one piece on the chessboard. Powerful. Yet limited without the rest of the set and a clear game plan. Large language models distill existing information. They lack authority on truth. Humans must stay in the loop as the final judge. (CIO)
But the core idea predates the ChatGPT boom. Evergreen’s 2022 book Autonomous Transformation: Creating a More Human Future in the Era of AI laid groundwork. It argues leaders should stop asking how to cut costs or speed existing processes. They should ask what new value they can create that customers will pay for. Netflix offers the classic case. The company moved from mailing DVDs to streaming. The shift created an entirely different experience. AI tools can accelerate such moves. They cannot define the destination.
Recent conversations reinforce the point. In a May 2026 podcast appearance, Evergreen contrasted adding a new piece to the board with rewriting the whole game. Most AI projects do the former. Few attempt the latter. A LinkedIn post from earlier this year put it simply: “There’s no such thing as an ‘AI Strategy.’ There is only a strategy. A strategy is a map from where you are to where you want to go.”
Executives who treat AI as the strategy often end up with automation theater. They deploy chatbots or agents that handle tasks. Customer satisfaction drops. They hire humans back. Klarna’s well-known experiment offers a cautionary tale. The company replaced hundreds of support staff with AI. Service issues followed. Human expertise returned to the front lines. Evergreen draws a clear lesson. Keep humans as the interface. Use AI as invisible middleware. Customers connect with people. Technology works behind the curtain.
But. Tasks and jobs are not the same. A job carries accountability for outcomes. AI cannot accept that responsibility. Leaders should separate repetitive tasks from roles that require judgment and relationships. Give the first to machines. Reserve the second for people. This decoupling creates space for higher-value work. It also demands clearer thinking about what work actually matters.
High agency accelerates progress. Evergreen’s own story proves it. He began at Accenture as a data-entry contractor. He taught himself SharePoint. He automated parts of his team’s workflow. Promotion followed. The episode shows what happens when individuals act without waiting for permission. Organizations that encourage this behavior move faster. Those that don’t stay stuck.
Current discussions echo these themes. A recent article on enterprise AI efforts in 2026 notes that many companies still announce ambitious visions yet struggle with execution. The gap persists. (Tomorrow’s Office) Another piece on strategic planning with AI describes the technology as an accelerator for scanning data, drafting objectives and tracking progress. Yet it cannot replace human judgment on priorities or trade-offs. (Brev)
Evergreen pushes further. He distinguishes problem solving from future solving. Problem solving removes what you don’t want. Future solving builds what you do want. The second approach opens more possibilities. It starts with a concrete picture of the desired state and works backward. Trends reported by analysts have value. They should not set the vision. Leaders own that task.
So what does effective practice look like? Define the future state in specific terms. Identify gaps between that state and current reality. Map the capabilities needed to close those gaps. Determine where AI, autonomous systems, real-time analytics or other technologies fit. Maintain human authority at critical decision points. Measure progress against the vision, not against activity metrics.
Recent X posts captured the sentiment well. One noted that most AI strategies amount to little more than vendor shortlists and adoption scorecards. Another emphasized that “10% more profitable” fails as a vision. Concrete, human-centered ambitions perform better.
Evergreen’s advisory work with Fortune 500 companies focuses on these shifts. He helps executives move from technology-led experiments to vision-led transformation. The difference shows in outcomes. Companies that begin with strategy make coherent choices about governance, talent, architecture and partnerships. Those that begin with tools accumulate incompatible systems and disappointed stakeholders.
Autonomous transformation, the concept at the heart of his book, aims higher than digital transformation. It seeks systems that augment human capability rather than replace it. The goal remains a more human future. Technology serves that end. Not the reverse.
Leaders who internalize this thinking ask different questions. Not “How do we use AI?” but “What future do we want to create, and how can these tools help us reach it?” The first question leads to fragmented projects. The second builds coherent advantage.
Plenty of organizations still chase the latest model or agent platform. They treat AI as the strategy. Results remain mixed. Evergreen’s record at Microsoft, his writing and his current advisory role suggest a different path. Start with vision. Build strategy around it. Deploy AI where it advances the map. Keep humans central. Create new value. The approach demands more upfront thought. It delivers more lasting impact.
That’s the choice facing executives in 2026. Follow the familiar pattern of use cases and incremental gains. Or do the harder work of defining a future worth building. The tools will be ready either way. Only one path turns them into advantage.
Why Vision Beats AI Tools: Brian Evergreen on Building Real Strategy first appeared on Web and IT News.