Research · October 2026
AI value comes from four organisational choices, not from more AI.
Almost every organisation now uses AI. Only 6% report significant value and at least 5% of EBIT from it (McKinsey, 2026). This synthesis sets out what that 6% does differently, drawing on studies from McKinsey, BCG, PwC, Gartner, MIT Sloan, NBER, IBM and others, with each source graded by strength of evidence.
Research synthesis by Mark Cameron, CEO, Alyve
Open the interactive research companionThe gap
use AI regularly in at least one function
report improved individual productivity
report any positive EBIT contribution from AI, essentially unchanged from 2025
are high performers attributing at least 5% of EBIT to AI with significant value, flat since 2025 (92 of 1,521 AI-using respondents)
McKinsey, The state of AI in 2026, August 2026
These figures come from one survey but measure different things, and should not be read as a single conversion funnel. Together they show the gap between use, individual benefit and enterprise economics.
The four choices
Concentrate on a few outcomes
The organisations creating value fund a small number of outcomes with milestones and money, and give everyone else the tools and permission to experiment and pass on what works.
Read the evidenceRedesign the work before automating it
Nearly three quarters of the leading-value cohort fundamentally redesign workflows because of AI, against about a quarter of others (McKinsey, 2026). Automating an unchanged process tends to move the bottleneck rather than remove it.
Read the evidenceMeasure results in dollars
AI reliably frees up time for individuals. It becomes value only when the organisation decides where that time goes and tracks what changes in dollars. High performers are twice as likely to have defined processes to measure impact (McKinsey, 2026).
Read the evidenceBuild a learning system
The leaders treat learning as infrastructure: time to experiment, shared standards for judging output, and a way to pass on what works. In a two-year field study, one organisation reached 141 organisation-wide AI solutions with structured support for experimentation; another ended with 3 after most participants disengaged (MIT Sloan, 2026).
Read the evidenceProof
In the three cases that worked, the work changed before the technology did. C.H. Robinson redesigned its freight workflows before automating them. JPMorgan Chase let everyone learn while putting its investment behind a few journeys. Moderna gave one owner the question of which work people do and which work technology does. Klarna is the cautionary case: when cost was the only measure, quality fell until customers noticed. All case figures are company-reported or from executive interviews.
Read the four casesAustralia
PwC’s 2026 study finds Australian enterprises lead the global sample on AI security, with 73% reporting robust protections against 69% of AI leaders. Only 7% have redesigned workflows around AI, against 56% of AI leaders. The governance foundation is built; the gap is in what gets built on top of it.
PwC Australia, 20 April 2026
See the Australian figuresThe question for your team: How long does a full turn of our learning loop take: weeks or quarters?
Inside the companion
The interactive companion holds the full argument and its evidence: the four cases in detail, ten questions that show whether an organisation is wide or deep, a 45-minute session plan for a leadership team, a field guide, and an evidence explorer with every source graded from A (experimental or longitudinal field evidence) to D (company-reported case).
How to read this
Most enterprise AI studies are observational surveys. Association is not causation, and different studies define value differently, so figures from different sources should not be combined as if they share one methodology. Evidence checked October 2026.
Evidence your audience can act on.
Mark turns this research into keynotes and facilitated sessions for conferences and leadership teams.
Explore Mark’s AI keynotes →