The adoption diagnosis
Wide, Not Deep
Access is not adoption. Activity is not transformation.
Definition
Wide, Not Deep describes organisations where AI use is widespread but remains shallow, individual and disconnected from operating systems.
The argument
The argument
The adoption story looks strong until the question changes from ‘who has access?’ to ‘what has changed?’
Why it matters now
The danger is mistaking individual experimentation for organisational capability while competitors redesign work, decisions and value creation.
Evidence
Alyve’s Corporate AI Adoption Intelligence (updated March 2026, drawing mainly on 2025 studies) placed McKinsey’s 88% reported-use figure beside BCG’s 2025 5% ‘future-built’ classification as a directional contrast—not a matched sample or common measure. BCG’s 2026 index shows more organisations now scaling, so the gap is narrowing for some.
Read the 88% / 5% methodology (2025 data) →The model
Depth appears when experiments connect to strategy, operating systems, capability and repeatable decisions.
Metaphor / example
A workforce can use copilots every day while approvals, incentives, job design and customer value remain untouched.
Leadership implication
Measure changed decisions and redesigned work, not licences, prompts or pilot counts.
Common misunderstanding
Wide use is not bad. It becomes a trap only when leaders treat breadth as proof of transformation.
External evidence and interpretation
What the research shows—and what Mark reads into it.
Published finding
BCG’s 2025 survey (1,250 senior executives) classified 5% of companies as ‘future-built’ and 35% as scaling AI. Its 2026 index (1,330 leaders) classified 7.5% as future-built and a further 41% as scaling—nearly half creating value. These are BCG’s own classifications, not a measure of all organisations.
Boston Consulting Group — Applied AI Index 2026 (press release, 30 September 2026) ↗Mark’s interpretation
Mark’s Wide, Not Deep diagnosis is about the distance between broad use and redesigned work. The 2026 data suggests that distance is closing for a growing group—which sharpens, rather than removes, the question of what has changed in your organisation.
Published finding
BCG’s 2025 release: 5% future-built, 35% scaling, based on 1,250 senior executives and AI decision-makers across nine industries.
Boston Consulting Group — The Widening AI Value Gap: Build for the Future 2025 (press release, 30 September 2025) ↗Mark’s interpretation
Dated context for the original synthesis; it does not mean ‘only 5% get any value’.
Questions people ask
What does Wide, Not Deep mean?+
AI use can be widespread while remaining shallow, individual and disconnected from the organisation’s strategy, decisions and operating systems.
Are the 88% and 5% figures from one study?+
No. 88% is McKinsey’s 2025 reported-use figure; 5% is BCG’s 2025 ‘future-built’ classification. They use different samples and definitions, so the pairing is a dated, directional contrast—not a matched measure. BCG’s 2026 index classifies 7.5% as future-built and a further 41% as scaling.
Sources
Boston Consulting Group — Applied AI Index 2026 (press release, 30 September 2026) ↗Boston Consulting Group — The Widening AI Value Gap: Build for the Future 2025 (press release, 30 September 2025) ↗Corporate AI Adoption Intelligence ↗Research library →Bring it into the room
Start a speaking enquiry →