The slide deck says 'responsible innovation.' The salespeople still type into CRM. The managers still spend half their weeks relaying information up a chain. The AI license is live; the old hierarchy is untouched.
That is the diagnosis Stephen Messer, co-founder of Collective[i] and Intelligence.com, lays out in a new Fortune commentary. He calls the pattern the 'AI Shuffle': the corporate habit of exchanging one technology logo for another while preserving every underlying assumption about how work gets done. 'It feels like progress because it generates activity,' Messer writes. 'It does not produce an advantage.'
Subtraction before automation
Messer's argument is structural, not motivational. The conventional corporate response to any new technology is addition — add a tool, add a dashboard, add a governance layer, add another system to the stack. The first instinct of an AI-first company, he contends, should be the opposite: subtraction.
His sequence is deliberate: question every requirement, remove unnecessary steps, simplify what remains, and only then automate. 'Automating a bad process does not make it a good process,' he writes. 'It makes the bad process faster, harder to see, and more expensive to unwind.'
He uses sales forecasting as the clearest case. For decades, companies have asked sellers to enter projections into CRM, managers to interpret them, and leadership to negotiate a number on a call that 'everyone knows is partly theater.' The data arrives late, incomplete, and distorted by incentives. The meeting exists, Messer argues, because the system cannot observe the buying process directly. The AI-era answer is not a more elegant forecasting meeting — it is a system that reads buyer behavior, market conditions, and relationship signals so that the ritual becomes unnecessary.
Software is not going down alone
The implications reach beyond any single workflow. Much of the enterprise software stack, Messer notes, was built to organize human data entry: applications that store records, route tasks, generate reports, and help managers reconstruct what happened after the fact. AI agents will increasingly observe activity, maintain context, and initiate or recommend the next best action — making large portions of that stack redundant.
Management layers built around information-gathering and translation across functions face the same pressure. Messer is careful to say leadership does not disappear; it transforms. The people who will matter most are what he calls builders — those who understand a real business problem, can apply technology to it, and are close enough to customers and operations to know whether the solution works. Those who lose relevance will be the ones whose roles depend on preserving friction, controlling access to information, or managing processes no one would design from scratch today.
He also warns that too many companies have placed their AI future in the hands of people selected to prevent mistakes rather than create new capabilities. Governance and security matter, he acknowledges — but a risk-averse posture applied uniformly to low-stakes experiments is itself a strategic liability.
What the numbers will eventually say
Messer does not offer a timeline or a market forecast. What he offers is a framework for measurement: the companies winning with AI do not track usage metrics. They track whether a specific business constraint has moved.
That is a standard free-enterprise investors should welcome. Capital rewards clear rules and measurable outcomes. A company that can demonstrate it has eliminated a cost center, compressed a decision cycle, or removed a layer of reconciliation work has something a slide deck cannot fake: margin. The AI Shuffle produces activity reports. Subtraction produces earnings. The market, in time, will price the difference.
