August 19, 2026 · 8 min read

What does aware, adaptive, alive actually mean for a company?

The Distance:
88% VS 39%

It means three capabilities that only work in sequence, and most companies buy the first and stop. McKinsey’s State of AI survey, published in November 2025 from 1,993 respondents across 105 nations, found 88 percent of organisations now use AI in at least one business function while only 39 percent report any effect on enterprise profit, and most of that 39 percent put the effect below five percent.

Aware is the ability to sense what is actually happening. Adaptive is the ability to change what you do because of what you sensed. Alive is when that loop runs on its own, continuously, without a person scheduling it. The distance between 88 and 39 is the distance between the first and the second. Almost nothing in the software market sells you the second.

88% / 39%
Use AI in at least one function, versus report any enterprise profit impact
McKinsey and Company, November 2025
67%
Of reported AI impact traces to organisational factors, roughly twice the share from individual mindset
Microsoft Work Trend Index, May 2026
66%
Of workers who save a day a week get little or no guidance on what to do with it
BCG AI at Work, June 2026
1%
Of capital is all that a third of companies reallocate from one year to the next
McKinsey and Company, August 2016
/ 01

Aware is the part everybody already bought

Sensing is now close to free. Dashboards, listening tools, intent data, session recordings, model driven summaries of everything that moved last week: any mid-market company can install more perception in a fortnight than a Fortune 500 had a decade ago. This is real progress, and it is also where the story usually ends.

The most quoted number in this argument deserves a careful look, because it is quoted badly. In July 2025, MIT’s Project NANDA published a preliminary report titled The GenAI Divide, by Aditya Challapally, Chris Pease, Ramesh Raskar and Pradyumna Chari, which stated that 95 percent of organisations are getting zero return from generative AI. Kevin Werbach of Wharton read the document repeatedly and could not locate the supporting analysis. His reading is that the underlying five percent figure in section 3.2 refers to a much narrower population, custom enterprise AI tools judged to have been successfully implemented, where success was defined as a marked and sustained productivity or profit and loss impact. Failing that bar is not the same as returning nothing, and 100 minus 5 is not a valid inference from a subset. Werbach asked NANDA to release the data or withdraw the claim.

Take the correction seriously and the honest version is less dramatic and more useful. The tools mostly work. What fails is absorption. McKinsey found that roughly two thirds of organisations have not begun scaling AI beyond isolated functions, and that its six percent of high performers were nearly three times as likely to have fundamentally redesigned individual workflows, with workflow redesign carrying one of the strongest contributions to business impact of the 31 variables tested. Microsoft’s 2026 Work Trend Index, drawn from 20,000 AI users across ten markets, puts it more bluntly still: organisational factors accounted for 67 percent of reported AI impact against 32 percent for individual mindset and behaviour.

THE STAIRCASE FROM ADOPTION TO ACTUAL RESULT 0255075100 88%USE AI IN ATLEAST ONE FUNCTION 33%HAVE BEGUNSCALING IT 39%REPORT ANYEBIT IMPACT 6%ARE HIGHPERFORMERS MOST OF THAT 39% PUT THE PROFIT EFFECT BELOW FIVE PERCENT
McKinsey and Company, The State of AI in 2025, published November 2025, 1,993 respondents across 105 nations.
/ 02

Adaptive is the part almost nobody built

Adaptation is not a feeling about change. It is a measurable willingness to move resources after you learn something. On that measure the record is poor and long standing. McKinsey’s research on resource reallocation, published by Yuval Atsmon in 2016, found that a third of companies shift just one percent of capital from one year to the next, with an average of eight percent, and that companies which reallocate actively delivered ten percent average total returns to shareholders against six percent for the sluggish. Ten years on, the finding still describes most boards we sit in front of.

The decision layer is no better. A McKinsey survey of more than 1,200 managers, published in 2019, found 61 percent said at least half the time they spent making decisions was used ineffectively, an amount the authors estimated at roughly 530,000 days of manager time a year at a typical large company. Sensing produces signal. Deciding is where signal is supposed to become action. If the decision layer is half wasted, adding perception simply increases the volume of things nobody acts on.

BCG’s June 2026 study of 11,749 workers across 14 markets shows exactly what this looks like at desk level. Seventy four percent of frontline employees are now regular AI users, up 23 points in a year. Forty two percent of them save a full working day or more per week. Sixty six percent receive little or no guidance about what to do with that time, and more than half do not redirect it into anything strategic. BCG’s phrase for the result is that the time saved leaks out of the organisation. In the same study, a clear strategy lifted measurable business impact by 25 percentage points while better tools alone moved it about five.

Microsoft’s data explains why so few companies close the gap even when they can see it. Sixty five percent of workers fear falling behind, and yet 45 percent said it feels safer to focus on current goals than to redesign how work is done, with only 13 percent reporting they would be rewarded for attempting reinvention if the results did not immediately land. Only 26 percent said leadership was clearly and consistently aligned on AI. Companies are not failing to adapt because they are unaware. They are failing to adapt because adapting is individually irrational in an organisation that only rewards the current plan.

THE LOOP, AND WHERE IT BREAKS IN MOST COMPANIES ALIVE THE WHOLE LOOP SENSE AWARE. 88% GOT HERE. DECIDE 61% CALL THIS TIME INEFFECTIVE ACT A THIRD REALLOCATE 1% OF CAPITAL A YEAR LEARN 66% GET NO GUIDANCE ON THE TIME THEY SAVED TEAL ARC: BUILT. GOLD ARC: WHERE THE LOOP GOES OPEN AND THE SIGNAL DIES.
McKinsey November 2025 and May 2019, BCG June 2026, McKinsey resource reallocation research 2016. Loop framing after Stafford Beer.

According to the cybernetician, the purpose of a system is what it does. Stafford Beer said that in 1979 and every quarterly review since has proved him right.

/ 03

Alive is what you can actually build this quarter

Alive is not a mood. It is the loop closing without supervision: sense, decide, act, learn, and back to sense. Ross Ashby set out the constraint in An Introduction to Cybernetics in 1956 with his law of requisite variety, which holds that a regulator can only absorb as much variety as its own capacity allows. A market that changes weekly cannot be governed by a company that decides annually. The mismatch is not a strategy problem. It is arithmetic.

/ 01

Name the sensor and the switch

For each of your five most important signals, write down who sees it, how often, and what specific decision it is allowed to change. If a signal cannot change anything, stop paying for it. Most companies discover they are buying perception with no attached authority.

/ 02

Move one percent on purpose

Pick a real number, five or ten percent of the marketing budget, and commit in advance to reallocating it every quarter based on evidence. Deciding the rule beforehand is what makes the reallocation survive the meeting.

/ 03

Book the saved hour

If a tool gives your team back four hours a week, put those hours in the calendar with a name on them before the tool arrives. Unclaimed time does not become strategy. It becomes more email.

/ 04

Redesign one workflow end to end

Not a pilot bolted onto an existing process. One complete path, from trigger to outcome, rebuilt for how it would work if you were starting today. This is the single behaviour that separates the six percent from everyone else.

/ 05

Reward the attempt that did not land

Publicly, once, in front of the team. Thirteen percent of workers believe reinvention is safe when results are uncertain. That number is set by what leadership visibly tolerates, and it can be changed in a single meeting.

/ 04

The part nobody wants to hear

Most companies will read all of this, agree with it, and change nothing, and the reason is not stupidity or bad intent. It is that the loop has no owner. Sensing has an owner, because somebody bought the tools. Acting has an owner, because somebody runs the campaigns. The connective tissue between them, the standing question of what we learned and what we will therefore do differently, sits in no job description, appears on no dashboard, and has no deadline. It belongs to whoever has ten free hours a week to think, and in a mid-market company that person does not exist.

Microsoft measured what happens meanwhile: employees interrupted every two minutes, 275 times a day, moving through 117 emails and 153 chat messages per working day. That is not an environment in which a company reflects on anything. It is an environment in which a company reacts, and reacting is the specific failure mode that looks most like being alive from the inside.

The stakes are not theoretical. Torsten Slok’s March 2026 update of the corporate longevity series puts the average tenure of a company in the S&P 500 at about 15 years, against roughly 33 in the 1960s. Companies are not dying of ignorance. They sensed it. They usually wrote a report about it. What they could not do was change what they did on a Tuesday because of what they learned on a Monday, which is the only definition of adaptive that has ever mattered. Aware is a purchase. Adaptive is a decision. Alive is a habit, and habits are the only one of the three that nobody can sell you. If none of it is written down anywhere, you do not have an operating system, you have a set of tools and a hope.

A company is alive to exactly the degree that something it learned this month can change what it does next month, and no software has ever supplied that on its own.

SOURCES: MCKINSEY AND COMPANY, THE STATE OF AI IN 2025, NOVEMBER 2025 · MIT PROJECT NANDA, THE GENAI DIVIDE, JULY 2025, AND THE WERBACH CRITIQUE, AUGUST 2025 · BCG AI AT WORK, JUNE 2026 · MICROSOFT WORK TREND INDEX, MAY 2026 AND JUNE 2025 · MCKINSEY ON RESOURCE REALLOCATION, AUGUST 2016 · MCKINSEY ON DECISION MAKING, MAY 2019 · APOLLO ACADEMY, TORSTEN SLOK, MARCH 2026 · STAFFORD BEER, THE HEART OF ENTERPRISE, 1979 · W. ROSS ASHBY, AN INTRODUCTION TO CYBERNETICS, 1956
DFJ
David Fox Jensen
Founder & Principal · The Sentient Marketer

20+ years leading marketing for 150+ brands. Now runs a principal-led fractional CMO and AI consulting practice for mid-market companies. Writes The Signal weekly on what the machines are learning about your business. Machines for scale, never for soul.

Subscribe

Wondering what the machines say about you?