Every strategy deck now has an AI slide, presented in a boardroom or reviewed by a five-person team around a laptop. Every organization has, at minimum, a Copilot license and someone unofficially in charge of figuring out what to do with it.
By that measure, the revolution is over and everyone won.
It isn’t, and they haven’t.
The technology itself is still evolving too, not just the organizations trying to use it.
Beneath the appearance of adoption sits a harder question that splits into three, and how you answer it says more about your company’s next five years than any AI roadmap on file.
- Are you convinced AI needs to be adopted, not as a talking point, but as something you would bet your operating model on?
- Are you confused about how much value it is actually providing, sensing that the spend and the results do not quite reconcile, but unable to say by how much?
- Or are you still hallucinating about whether the decision is even yours to make, treating indecision as a safe, neutral place to wait, when it is, in fact, a decision, with consequences of its own?
Most leaders, if they are honest, will find themselves in more than one.
Convinced
McKinsey’s State of AI in Business 2025 survey found that 88% of organizations now use AI regularly in at least one business function, up from 78% a year earlier.
If that were the whole story, this section would be short.
Only about 6% of respondents qualify as “AI high performers,” attributing 5% or more of their bottom line to AI. Just 39% report any bottom line impact at all.
Adoption is nearly universal. Conviction, measured in dollars rather than dashboards, is rare.
What separates that 6% is not model sophistication. It is the seriousness of their commitment. High performers are roughly three times more likely to have fundamentally redesigned workflows around AI instead of layering it on top of how things already worked.
They did not add AI to the org chart. They rebuilt parts of it.
JPMorgan Chase is the clearest public example: its LLM Suite reaches more than 230,000 employees, saving an estimated three to six hours per person a week on tasks like presentation drafting and financial document comparison. Few organizations can match that scale, but the discipline behind it, measuring hours actually returned rather than licenses issued, scales down to any size.
That is what conviction costs. Not a pilot budget, not a task force, but a multi-year commitment, a willingness to redesign how work gets done, and the discipline to measure results in bottom line impact rather than adoption metrics. Very few organizations have made that trade, and fewer still are honest with themselves about which side of it they are on.
So, the question is not whether your company uses AI. It is whether you could attribute a specific, defensible percentage of your bottom line to it.
If the honest answer is “not yet,” you are not in the convinced 6%. You are somewhere else.
Confused
If conviction is rare, confusion is common, and more forgivable than most leaders think.
You ran the pilots, and now you are staring at a number that does not reconcile with the spend. That is not failure to act. It is failure to see a return, and it turns out you have a great deal of company.
MIT’s NANDA initiative found that despite $30 to $40 billion in enterprise generative AI spending, 95% of organizations are getting zero measurable return and just 5% of pilots extract real value.
The divide, researchers found, is determined by approach, not model quality or regulation.
That distinction matters. The confusion is not a verdict on whether AI works. It is a verdict on how it was implemented.
Most enterprise tools are generic and brittle in real workflows. They cannot retain feedback or adapt with use. In over 90% of firms, employees quietly route around failed official pilots using their own personal AI tools instead.
Sit with that for a moment. The value is real, it’s just happening in tools your company never sanctioned or measured.
This is where the confused and the convinced diverge, and it is not about budget. Organizations that cross into the high performing tier treat a disappointing pilot as diagnostic information rather than a verdict, and rebuild the workflow instead of asking the tool to fit an unchanged one.
So, ask yourself plainly: is it the tool, the workflow, or the data that is failing?
If you cannot answer with precision, you are not confused about AI. You are confused about your own organization, and AI simply made that visible.
Hallucinating
There is a third position, and it is the one leaders talk about least, because it does not feel like a decision at all. It feels like patience. You have not rejected AI, and you have not committed to it either. That feels safe, but its not, and “hallucinating” applies here in two ways at once.
The first is figurative: treating indecision as a neutral holding pattern is itself a kind of hallucination. There is no neutral position. While you wait, your workforce and your competitors are not waiting with you.
The second is literal, and it is where the consequences of the first tend to show up. When organizations avoid a deliberate, governed decision about AI, the decision does not disappear. It gets made anyway, later, by someone under pressure, with no framework in place.
The legal profession shows this clearly, because courts write down what happened. Legal analytics tracking now counts more than 700 court cases involving AI generated hallucinations or fabricated content submitted as fact, with sanctions that have escalated roughly elevenfold since 2023.
The pattern has nothing to do with the sophistication of the technology. It is almost never a governance framework failing under pressure.
It is the absence of one entirely, because the decision to build it was never made.
The hallucination is not really the model’s failure. It is the organization’s unmade decision, arriving late and in public.
So ask it outright: has your organization actually decided what responsible AI use looks like, or are you simply hoping the question does not come up?
If you have to think about it, you haven’t decided. You’ve been drifting, and calling it patience.
The Wrap
None of these three questions has a single answer for your entire company. You can be convinced about how you handle your finances, confused about your marketing spend, and still hallucinating about anything that touches your legal exposure.
Most organizations, whatever their size, are all three at once, just in different corners of the business.
That is useful: the work is not one company wide verdict, but a room-by-room audit.
- Where have you redesigned how work happens, and can you prove it in your bottom line?
- Where have you piloted without knowing why it did not land?
- And where have you simply avoided deciding?
The revolution does not wait for leadership teams to feel ready. It rewards the ones who can answer these three questions honestly before someone else forces the answer on them.
