AI Investments Are Not All The Same

Part 1: What I Learned the Hard Way in My First Board Meeting as a Technology Leader

Part 2: What I Learned Sitting Across from the Sharpest People in the Room

In the first two parts of this series, I wrote about learning to show up effectively in the boardroom as a technology leader. The hard lessons around context, credibility, and what boards actually care about.

This third part is about a conversation happening in nearly every boardroom right now. One that most technology leaders are not yet framing well.

AI investment.

The problem I keep seeing is that we treat all AI investments the same. Everything lands in the budget review as one undifferentiated list. A dollar amount, a timeline, an expected benefit. And the wrong questions get asked. Or no questions get asked at all.

I use a simple framework I call E³: Explore, Experiment, Execute.

If your board can walk out remembering three words, they can ask better questions and hold the right people accountable.

Each phase is different. Different risk, different standard of approval, and different conversation.

Explore

Exploration is about developing judgment.

You are not buying AI at this stage, you are buying informed perspective.

The goal is to understand what AI can actually do in the context of your business, including your data, your workflows, and your competitive environment. The outputs are observations, prioritized opportunities, and honest assessments of where you are not ready.

The investments here are modest in dollars but significant in leadership time.

Explore investments do not need ROI projections. They need strategic rationale and executive sponsorship.

Justification lives at the executive team level. The CIO or CTO, working with the CEO, should be able to answer three questions.

• What are we trying to learn?

• What are we assessing?

• What would cause us to move forward, or stop?

Board visibility at this stage is light. They need to know leadership has a coherent AI learning strategy. Not a free-for-all, or paralysis, just a clear and disciplined approach.

Experiment

Experimentation is where you move from judgment to evidence.

You have identified use cases that look promising. Now you are testing whether they actually work in your environment, with your real data, against your real constraints.

This is the phase most organizations are in right now. And the phase most prone to confusion.

Experiments get mistaken for projects, but they are not the same thing. A project has a delivery commitment, while an experiment has a learning commitment.

The question an experiment has to answer is not “did we build it?” It is “did it work well enough to scale?”

Justification here shifts from strategic rationale to hypothesis and measurement.

• What are you testing?

• How will you know it worked?

• What is the minimum result that justifies moving forward?

The executive team owns the gate decisions. This is where results either earn the right to advance, or they don’t.

The board should see the portfolio of active bets and the governance structure behind them. They want to know someone is accountable for making the go or no-go calls.

Execute

Execution is where AI moves from initiative to operation. You have learned what works. Now you are committing to build it, scale it, and run it.

This is a different obligation entirely.

Architecture decisions have long tails, and vendor contracts have real terms. Change management, training, and operational support all become line items. The connection to business outcomes has to be explicit.

At the Execute stage, the burden of proof is no longer whether AI can do this. It is whether your organization can actually realize the value.

One practical tool here is a structured process map of your enterprise. A framework like the Macro Process Wheel gives leadership a common language for identifying which processes AI will touch, how the work inside them will change, and what dependencies that creates.

The value is precision. Instead of claiming AI will improve operations, you can point to specific processes, specific changes, and specific owners. That is what turns a promising experiment into a credible execution plan.

Justification at this phase is a full business case. Investment cost, value timeline, risk factors, and the organizational changes required to capture the benefit.

The board becomes a direct stakeholder here. These are material investments with material expectations. A board that only hears about AI after it is already operational has been underserved.

Where the Justification Lives

Here is the simple version.

• Explore: Justification lives with the executive team. Board visibility is awareness only.

• Experiment: Justification lives with the executive team. The board sees the portfolio and the governance model behind it.

• Execute: Justification requires a full business case reviewed at the executive level. The board is a direct stakeholder and holds management accountable for results.

The Conversation Worth Having

The organizations that skip Explore and Experiment and rush straight to Execute often build the wrong things at scale. They commit to use cases that were never validated, and they find out after they have already paid for it.

The organizations that never leave Explore and Experiment end up with impressive pilot portfolios and no business impact. Boards lose patience, as they should.

The best technology leaders I know treat Explore like a learning cost, experiment like a bounded bet, and Execute like a capital commitment that requires the same rigor as any other major investment.

Not all AI investments are the same. The organizations that govern them like they know the difference are the ones that will actually capture the value.