“Go Use AI” Is Not a Strategy

People are already using AI at work.

Microsoft and LinkedIn’s 2024 Work Trend Index found that 75% of knowledge workers surveyed were using AI at work. More importantly, 78% of AI users reported bringing their own AI tools into the workplace rather than relying exclusively on tools provided by their employers.^[1] Among small and midsized companies, that figure was 80%.^[2]

That tells us something important.

The problem for many companies is no longer getting employees interested in AI.

The problem is turning that interest into business value.

An employee discovers that ChatGPT can improve an email. Someone in marketing uses AI to brainstorm headlines. A salesperson summarizes meeting notes. Someone else creates a presentation faster.

Those can all be useful.

But twenty people individually saving fifteen minutes does not automatically add up to an AI strategy.

The question leadership should be asking is not:

“Are our people using AI?”

It is:

“Where can AI measurably improve how our business operates?”

Those are very different questions.

One of the easiest mistakes to make with AI is starting with the technology.

“We bought Copilot. What can we do with it?”

“We have ChatGPT. Where should we use it?”

“What could we automate with AI?”

Turn the question around.

Start with the business.

• Where are people spending too much time?
• Where are customers waiting?
• Where are employees repeatedly searching for information?
• Where are people copying information from one system into another?
• Where are reports, proposals, estimates, summaries, or other documents being created manually?
• Where are errors expensive?
• Where does valuable knowledge exist primarily inside the heads of a few experienced employees?

Those are potential AI opportunities.

Now AI has a job to do.

That distinction matters because research increasingly suggests that business value comes from changing work, not simply distributing AI tools.

McKinsey’s 2025 global AI survey examined 25 organizational attributes associated with value from generative AI. Of those attributes, redesigning workflows had the strongest relationship with an organization’s ability to report EBIT impact from generative AI. Yet only 21% of respondents whose organizations used generative AI said they had fundamentally redesigned at least some workflows.^[3]

Watch this.

That gap is important.

Companies are adopting AI much faster than they are redesigning the work around it.

There is good evidence that AI can improve performance when it is applied to a specific job.

Researchers studying 5,179 customer-support agents found that access to a generative AI assistant increased issues resolved per hour by an average of 14%. The improvement was substantially greater for less experienced and lower-skilled workers.^[4]

Notice what happened there.

The company did not simply tell 5,179 employees to “go use AI.”

AI was applied to a defined workflow with a defined purpose, and its impact could be measured.

That is the model businesses should pay attention to.

Not:

Tool → Employees → Hope

But:

Business problem → Workflow → AI application → Measurement → Improvement

That is the beginning of an AI strategy.

You do not need a 75-page AI transformation plan before anyone experiments.

In fact, that can be just as counterproductive as having no strategy at all.

A smaller business can begin with something much simpler.

1. Pick one business problem worth solving.
2. Define what happens today.
3. Establish a baseline.
4. Choose one place where AI might improve the process.
5. Run a controlled experiment.
6. Measure what changed.

If it works, improve it and expand it.

If it doesn’t, learn from it and try something else.

Suppose preparing a customer proposal currently takes three hours.

That gives you something concrete to work with.

Could AI safely reduce the research, drafting, or review time?

Could three hours become two?

Could two become one?

And just as importantly, does the resulting proposal remain accurate and useful?

Now you can measure something.

Time saved. Cost reduced. Response time improved. Errors prevented. Revenue increased. Customer experience improved.

Worth measuring.

Those are business outcomes.

“Number of employees using AI” is not.

None of this means employees should stop experimenting.

Quite the opposite.

The people doing the work often understand its inefficiencies better than anyone else. They are likely to discover valuable AI applications that management would never identify from a conference room.

Give them permission to explore.

But give that exploration some direction.

Ask employees:

Then listen.

Those answers can become the beginning of your AI opportunity list.

From there, leadership can prioritize opportunities based on business value, feasibility, risk, and measurable impact.

That is very different from simply buying licenses and hoping innovation happens.

AI adoption is not the objective.

Better business performance is.

Your employees using AI may be encouraging. It may even produce some immediate productivity improvements.

But activity and progress are not the same thing.

The companies that get meaningful value from AI will be the ones that connect the technology to actual business problems, redesign the work where necessary, establish reasonable safeguards, and measure what changes.

So yes, tell your people to explore AI.

Just don’t stop there.

Give them a problem worth solving.