The Roadmap: Tie Everything to Value
Capability one, and the one that quietly governs the rest. McKinsey is blunt: a digital roadmap is not a list of technologies. It is a plan to capture specific business value in specific places.
A digital roadmap that lists technologies is just a shopping list. A real one names the value it's going to capture, and where.
This is capability one, and it quietly rules the other five. The trap is the roadmap that reads like a technology catalogue — “adopt the platform, roll out the model, migrate the data.” Every item is real, and the list still tells you nothing about why, or what changes if you finish it. A value-led roadmap inverts that: it starts from a specific part of the business and the specific value to be won there, then works backwards to the technology that earns its place.
Technology-led — what fails
“Adopt the platform, ship the model, migrate the data.” A list pointing nowhere — finish it and no one can say what changed.
Value-led — what good looks like
“In claims, capture £12m by cutting cycle time in half.” The figure decides what gets built, and in what order.
The unit that makes this concrete is the business domain: a coherent slice of the company where value lives — a customer journey, a product line, an operational process like claims or fulfilment. You don't transform “the company.” You pick a domain, size the prize in it honestly, and aim the six capabilities at that prize. Once the prize is named, the order stops being a matter of taste: you sequence domains by value at stake against effort to capture — biggest, most reachable prizes first. That ordering is unglamorous on purpose, and following it anyway is most of the discipline.
Why this matters more in 2026, not less
The gen-AI years made the value discipline sharper, not softer. Surveys now show the uncomfortable pattern bluntly: around eight in ten companies have deployed generative AI, and about the same share report no material impact on earnings. The technology spread; the value didn't — because so much of the effort was never tied to a specific number in a specific place. The roadmap is the capability that closes that gap. It's also where the horizontal-versus-vertical choice lives: a broad copilot rolled out to everyone scales easily but moves little money, while value tends to sit in vertical use cases wired deep into one workflow. Lead with value and you aim at the vertical prize on purpose.
Where it goes wrong
Falling in love with the technology and reverse-justifying the value afterward — you can always invent a business case for a thing you've already decided to build. The tell is the order you did it in. The second miss is staying abstract: “improve customer experience” is not a domain with a number. If you can't put a figure and a place on it, you don't have a roadmap — you have a wish.
Try this
Pick one initiative you're planning and write a single sentence: “In [domain], we will capture [specific value] by [doing X].” If you can't fill the value blank with something concrete, you've found the real problem — and you've found it before spending the budget, which is the entire point of leading with value.
Grounded in Lamarre, Smaje & Zemmel, Rewired (McKinsey), and McKinsey's gen-AI “value paradox” research (The state of AI, Seizing the agentic AI advantage).