The Two AI Conversations Happening at Once
By Jeff Chung - Chief Strategy Officer
05/13/2026
Two conversations are happening about AI right now, often in the same room.
In one, AI is transforming everything. Productivity is up across every function. Drafts get written faster. Reports get assembled faster. Companies are reorganizing around what AI can do, the future is rewriting itself, and the people not paying attention will be left behind.
In the other, AI is mostly failing to deliver. Pilots are getting cancelled. Boards are asking where the return is. Independent research keeps showing that the share of organizations seeing real bottom-line impact from AI is small, sometimes very small. The hype was hype.
Both conversations cite credible numbers. Both have real evidence. Both are happening in the same boardrooms, sometimes in the same people.
They are both, in their honest moments, partly right. The trouble is neither one is asking the question that actually matters.
Both sides have honest evidence
The transformation camp is measuring output. When someone says AI is changing everything, they are typically pointing at speed of work. Output that used to take a person an hour now takes ten minutes. Output that used to take a team a week now takes a day. That movement is real, repeatable, and visible at every level. The vast majority of organizations actively investing in AI report some level of productivity gain. That tracks with what almost any team using AI day-to-day will tell you.
The skeptic camp is measuring outcomes. When someone says AI is not delivering, they are typically pointing at financial performance. Has AI shown up in EBIT? Has it actually moved the line on cost, on revenue, on margin? At the company level, the data is much harder to find. Independent research consistently puts the share of organizations seeing meaningful financial impact from AI in single digits or low double digits, depending on how the question is asked.
Both pictures are honest. They are answering different questions.
Output is one thing. Outcomes are another. They are not the same.
"Is AI working?" is the wrong question
It is tempting to think one side is right and the other is wrong, that the truth must lie somewhere in the middle, that with enough patience the camps will resolve themselves. They probably will not. Because they are both answering "is AI working?" when the more useful question is something else entirely.
The question that gets you somewhere different is this. Is AI changing what work is?
If AI is just making existing work cheaper, that is exactly what the data shows. Productivity goes up at the individual level. Outcomes mostly do not move at the company level. Output is cheaper. The work itself, the steps and handoffs and decisions and approvals, is mostly the same shape it was eighteen months ago. Output gets faster. Outcomes do not follow.
That is not a failure of AI. It is also not a victory. It is a description of what most organizations have asked AI to do, which is to make existing work faster without changing what existing work is.
Where the actually interesting movement is
There is a small group of companies in a different conversation entirely. They are not asking whether AI is working or whether it is failing. They are asking a question that sounds simpler than it is.
Which of our people are already doing work that doesn't match their title?
That question matters because the work that doesn't match the title is usually the work AI is most ready to take on or transform. The drafting that happens before the writing. The routing that happens before the reviewing. The first-pass analysis that happens before the actual analysis. The meeting prep that happens before the meeting. None of it shows up on an org chart. All of it is real work, and a lot of it is exactly what AI is good at.
Companies looking at this seriously start asking what their people would actually do if those slices of work disappeared. Which becomes a different question again, because suddenly it is not "how do we use AI?" but "what does our work need to be, given what AI can now do?"
That is the conversation that actually moves something.
Where this goes
Most boardrooms right now are still picking sides between the transformation camp and the skeptic camp. The debate feels important. It is mostly an old debate in new clothes.
The companies that will look strategic in eighteen months are not the ones who picked correctly between hype and doubt. They are the ones who quietly stopped having the debate and started asking what work is actually for, given the new capabilities sitting on every desk in the building.
The two conversations will keep going. They will keep producing contradictory headlines. Both sides will keep being half-right.
The question worth asking is happening somewhere else.
Chief Strategy Officer, Arkane Digital
Jeff advises organizations on AI transformation, focusing on connecting business strategy to practical implementation without fragmented or reactive adoption.