The 7 Blinds of AI Adoption that Boards Fail to See

Many boards are accelerating AI adoption as if slow adoption will mean big losses.  Truth be told, without understanding the governance and human resource implications, boards are overestimating AI benefits and ignoring the risks of AI adoption. 

For example, one public accounting firm uses AI to analyse clients' financial statements and did not realize that such data entered into the internet is available for all to see.  This creates a big privacy issue that breaks trust with its clients. 

The question is: Are we actually ready to use AI responsibly and productively?

Buying AI tools is easy. Readiness requires clean and accessible data, appropriate infrastructure, capable people, redesigned processes, cybersecurity, controls and management accountability.

A company can be technologically AI-enabled but organisationally AI-unready.

2. Underestimating the human problem

The biggest barrier may not be technology. It may be people.

Employees may fear replacement. Managers may not know how to redesign work. Experienced staff may resist new processes, while enthusiastic employees may use AI without sufficient controls.

Boards should therefore ask:

Which jobs will change? Which skills become more valuable? What skills disappear? Who needs reskilling? And how will we bring our people with us?

AI transformation without talent transformation is unlikely to deliver its promised value.

3. Not knowing where AI is already being used

This is potentially more dangerous than slow adoption. Employees may already be putting customer information, company documents, financial data or intellectual property into public AI systems.

Management may therefore tell the board:

“We haven't implemented generative AI yet.”

while hundreds of employees are already using it.

The question should be:

“Where is AI being used in our organisation—with or without our approval?”

That includes third-party suppliers. Your company may not be using AI directly, but an important vendor may be using it to deliver services to you.

4. Trusting AI because the answer looks convincing

AI can produce an answer that is articulate, confident—and wrong.

That creates a new governance danger.

When AI produces a financial analysis, credit recommendation, legal summary, hiring shortlist or customer response, who is accountable for the decision?

“We followed the AI recommendation” cannot become the corporate equivalent of:

“The computer says so.”

Boards need clear human accountability, especially for high-impact decisions.

5. Focusing on AI risks but forgetting AI opportunity risk

Governance should not become the department of “No.”

There are risks in adopting AI too quickly—but there are also risks in moving too slowly.

What happens if competitors reduce costs by 30%?

What if they serve customers twice as quickly?

What if they develop products in half the time?

So boards should consider both:

Risk of AI adoption

and

Risk of AI non-adoption.

A good board helps management determine where to accelerate, where to experiment and where to apply the brakes.

6. Measuring activity instead of value

A company can announce 50 AI projects and still create very little value.

Boards should stop being impressed by:

“We trained 5,000 employees.”
“We launched 30 AI pilots.”
“70% of employees now use AI.”

Those are activity measures.

Ask instead:

What revenue did AI create?

What costs did it reduce?

What decisions improved?

What customer problem did it solve?

Which risks increased as a result?

AI adoption is not the objective.

Sustainable value creation is.

7. Failing to imagine the second-order consequences

This may be the biggest blind spot.

Suppose an AI project works spectacularly well.

Then what? If AI removes 30% of routine work, what happens to junior employees who traditionally learned through that work?

• If customers increasingly interact with AI agents, what happens to human relationships?
• If everyone relies on the same AI models, does the company's thinking become less differentiated?
• If AI makes decisions faster, can governance and controls keep up?

And if humans gradually stop exercising judgment because AI usually gets things right, what happens when AI gets the important one wrong?

These are not merely technology questions.

They are board questions.

The question I would put to every board

Instead of asking management:

“What is our AI strategy?”

I would ask:

“How will AI change our business model, people, risks and competitive advantage, and what must this board do differently as a result?”

Then I would follow with four deceptively simple questions:

• What are we trying to achieve with AI?
• What could go wrong?
• What are we not ready for?
• Who is accountable when it does go wrong?

This connects directly with our Brakes to Go Faster philosophy.

A board's job is not to stop AI.

Neither should it jump onto the AI bandwagon simply because everybody else is accelerating.

The faster companies accelerate with AI, the stronger their governance needs to become.

AI is the new engine.

Governance provides the brakes, steering and dashboard.

And perhaps the greatest risk for boards today isn't simply moving too fast with AI.

It is moving fast without knowing whether the organisation is ready, where it is going, or who has their hands on the wheel.