The AI value gap is a management problem

In short: nearly every organization now uses AI, but only about a third report any EBIT contribution and only about 6 percent capture significant value. The difference is management: a single owner of the AI portfolio, pilots designed to measure, work redesigned around the tool, and a named person who converts saved time into a result.^1,2

Adoption is no longer the question. In McKinsey’s 2026 global survey (1,719 respondents, fielded in May and June 2026), nearly nine in ten respondents reported regular AI use in at least one business function, and 40 percent of those at organizations with more than $1 billion in revenue said they were scaling AI agents, up from 27 percent a year earlier.¹

Value is the question. In the same survey 37 percent said AI had contributed to their organization’s EBIT, essentially unchanged from 2025, and only about 6 percent qualify as high performers that attribute at least 5 percent of EBIT to AI. Fewer than half, 44 percent, reported AI scaling across the enterprise.¹

Three studies, one direction

Boston Consulting Group’s 2025 research sorts companies into three groups: about 5 percent “future-built” leaders capturing outsized value, about 35 percent scaling and beginning to see value, and about 60 percent reporting minimal revenue or cost gains.² MIT’s NANDA initiative drew a harsher line from its 2025 sample: roughly 95 percent of generative AI pilots delivered no measurable profit-and-loss impact.³ That figure is best read as directional, since the sample was small and the definition of “pilot” broad. The direction, confirmed three times, is the point: most organizations are using AI, and most are not yet being paid for it.

What the leaders do differently

The differences are management variables, not technology. In BCG’s research, nearly 100 percent of future-built companies report deeply engaged C-suites, against 8 percent of laggards, and some 62 percent of their AI initiatives have already been deployed, against 12 percent for laggards; and leaders concentrate investment on a few high-value areas rather than spreading it thin.² McKinsey associates enterprise impact with redesigning workflows around AI rather than bolting tools onto existing ones, with senior ownership, and with scaling beyond pilots.¹ MIT attributes most pilot failures to a learning gap between tools and organizations.³

In practice four causes recur: no single owner of the AI portfolio; pilots designed to impress rather than to measure; tools bolted onto work instead of built into it; and unmanaged risk, which ends programs faster than weak business cases do.

Where value is created, and where it is captured

AI creates value at the task: a letter drafted in five minutes instead of twenty. The organization captures value only when someone does something with the difference: more volume with the same team, time redeployed to higher-value work, less overtime, a quality metric improved, a cycle shortened that a customer notices. Minutes saved across a thousand people become money only when a manager converts them, and that conversion is a management act with a name, an owner and a number.

The figures in the right-hand column are illustrations of the form, not benchmarks; the point is that each names a result, an owner and a measure. The simplest discipline we know: every use case states, in one sentence at the top, how its value is captured. “Value is captured when…” followed by the metric and the owner. If the sentence cannot be written, the use case is not ready to fund.

Accountability comes with the use

The other half of management is accountability. In 2024 British Columbia’s Civil Resolution Tribunal rejected Air Canada’s argument that its website chatbot was “a separate legal entity that is responsible for its own actions” and held the airline liable for what the chatbot had told a customer, awarding the fare difference as damages.⁴ The award was modest, but the reasoning carries over: an organization answers for what its AI systems say and do. Ownership of value and ownership of risk are the same job, and it is the job the gap is waiting for.

Earlier briefings fill in the method: why self-reported productivity numbers mislead, how JPMorgan measured an assistant’s spread, and last year’s use-versus-impact survey. Building the use-case portfolio and its value cases is the subject of Module 3 of the AIMP program, and measuring the result is Module 11.