The Doing Gap

Artificial intelligence is now inside almost every serious organisation on the planet. The tools are bought, the licences are issued, the all-hands presentations have been delivered. By most measures, the adoption problem is solved.

So why aren't the results showing up?

That is the question sitting quietly at the back of every board meeting right now. Not "should we be doing AI?" That conversation is over. The question is why so much investment, activity, and genuine employee effort is producing so little that shows up on the bottom line.

A landmark study makes the problem impossible to ignore. The National Bureau of Economic Research surveyed nearly 6,000 senior executives across the US, UK, Germany, and Australia and found that nine in ten firms report no measurable impact from AI on either employment or productivity over the past three years. Not a marginal improvement. Nothing. And these are organisations that are actively using it.

That is the most important number in business right now. Not because it means AI does not work. It means most organisations are using it in a way that does not work.

Here is what is happening. When an employee uses AI to do their own job faster, that time saving stays with them. It does not automatically move to the customer, the product, the service, or the margin. The work gets done more quickly, but if the process around that work has not changed, the business captures almost none of the gain. The individual is more productive. The organisation is not.

Think about what this looks like in practice. A procurement manager at a mid-sized manufacturer uses AI to analyse supplier contracts in a fraction of the time it used to take. She flags risks faster, builds reports more efficiently, and leaves the office earlier. Real improvement. But the approval process she feeds into still has five sign-offs. The supplier relationships are still managed the same way. The savings she could be unlocking through better negotiation are still sitting on the table. Her productivity gain evaporated at the organisational boundary.

That is not a technology problem. It is a design problem.

The organisations pulling ahead have understood something the majority are still missing. They are not asking how AI can help their people work faster. They are asking which parts of the business should work completely differently now that AI exists. That is a different question, and it leads to different decisions.

Singapore's Government Technology Agency embedded AI directly into public servant workflows, not as an optional tool sitting beside the existing process, but as a redesigned way of working where AI handled the administrative load and officers focused on judgment calls that required human thinking. The result was administrative time cut nearly in half. That did not come from giving people access to AI. It came from redesigning the work itself.

A financial services company rebuilt its meeting and follow-up process around AI, automatically capturing actions from video calls, drafting follow-through reminders, and tracking commitments across the organisation. The value was not that individuals saved time composing emails. It was that the organisation stopped losing decisions in the space between meetings. Execution improved because the system improved, not just the people in it.

The pattern is consistent across sectors and borders. The organisations reporting genuine commercial results from AI are not the ones with the highest adoption rates or the biggest budgets. According to Deloitte's 2026 enterprise research, only 34% of organisations are genuinely reimagining their business with AI. The remaining two-thirds are optimising, layering AI tools onto processes that remain fundamentally unchanged.

Optimising is not nothing. But it is not defensible, either. A competitor who redesigns their service delivery or cost structure around AI does not just become more efficient. They become structurally cheaper, faster, or better in ways that are difficult to close once the lead is established.

The cost of staying in optimisation mode is not abstract. It is the contract that goes to the competitor who can turn around a proposal in hours rather than days. It is the government service that citizens abandon because the private sector equivalent is faster and more responsive. It is the talent that leaves for the organisation where AI is making their job genuinely better, not just adding new tools to an old job description.

None of this requires being an AI expert. It requires being honest about one question. Are we using AI to do the same things faster, or are we using it to do different things altogether?

Most organisations, if they answer that question plainly, know they are still in the first camp.

What to do about it

Pick one process that is genuinely broken, slow, or expensive and redesign it with AI at the centre. Not a pilot sitting alongside the existing operation. A proper redesign with a clear owner and a measurable outcome. One well-executed change will teach more than a hundred licence deployments.

Ask where the time savings are going. If your people are using AI and getting faster, find out what is happening with that recovered time. If it is being absorbed by the same old workflow, the workflow is the problem. Fix that before adding more tools.

Separate individual productivity from organisational productivity in how you measure progress. If your AI metrics only capture usage and individual efficiency, you are measuring the wrong thing. The question is whether the business is faster, cheaper, or better. That requires different data.

Find the redesigners. In most organisations there are one or two people who have quietly restructured how their team works around AI and are producing results that stand out. Find them. Understand exactly what they changed. Then ask whether it can be replicated more broadly, because it usually can.

Stop counting how many people are using AI. Start counting what has changed because of it.

That is the shift. And the businesses that make it in the next twelve months will not just be ahead. They will be in a different race entirely.

𝛀

If you found this Independent A.I. Think Tank release useful, consider subscribing to D!sruptive newsletter. Built for business leaders who need to cut through the noise and understand what AI means for their organisation.

If you need the help discussed in this article, get in touch to see how we can work with your business 🇺🇸 +1 512 518 4196, 🇦🇪+971 585 629 975 or hello@byDisrupt.com