Most AI still stops at the pilot

Almost every company is trying AI, but very few have made it part of how the whole business runs. In McKinsey's 2026 survey of 1,000 executives, close to 90 percent of organisations said they were at least experimenting with AI, yet only 7 percent had scaled it across the enterprise.

The gap is rarely about the model. It is about whether the business has the habits that let a new tool change daily work: clear goals, shared numbers and people who act on them.

Two forces that feed each other

The same research found that companies using AI across many functions earned close to double the profit margin of companies using it in only a few teams. Companies with mature operating practices showed a similar lift.

The two reinforce each other. Strong operations make AI easier to roll out, and AI that is built into daily work makes operations sharper. Treat them as one programme, not two.

  • Operational excellence gives AI clean data, clear owners and a place to land.
  • AI gives operations faster decisions, earlier warnings and less manual effort.

What strong operations look like in practice

Operational excellence sounds abstract, but it shows up as a short list of everyday habits. These are the ones that separated the most profitable companies in the survey:

  • A few clear KPIs that cascade from leadership down to the front line.
  • Performance measures that connect daily work to the company's purpose.
  • Real-time data used to move people and budget to where they matter most.
  • Customer feedback that flows back into the process continuously, not once a year.

If these habits are missing, an AI tool has nothing solid to plug into. Fixing them first is often the fastest route to AI results.

Build the loop in six moves

  • Pick a few operational outcomes, such as throughput, yield, service level or response time, instead of funding dozens of pilots.
  • Choose first use cases that cross teams and sit inside core workflows, so they can scale.
  • Build the data, performance tracking and governance alongside the technology, not after it.
  • Put technical people next to the operators who will use the tool every day.
  • Match the pace of change to what the organisation can absorb.
  • Use operational discipline as the stabiliser and AI as the accelerator.

A note for mid-sized companies

The research found that mid-sized companies, roughly 500 to 5,000 employees, are often squeezed: they lack the scale of the largest firms and the agility of the smallest. They can still win by focusing on one or two high-value processes, getting the operating basics right there, and expanding from proven results.

Source

Figures in this guide come from McKinsey & Company, "Putting AI to work: The operational excellence imperative" (2026). The guidance is Aarvik Digital's own.