Picture a board meeting about artificial intelligence.
One person presents a tool. Another suggests a training program. Someone asks how much it would cost to automate part of customer service. The conversation moves on to vendors, licenses and timelines.
Until a question comes up:
What do we want to become possible in our business?
Answering requires looking beyond the tasks that already exist. It may involve an experience the company cannot offer yet, a decision that arrives too late or an opportunity that always seemed to require an unworkable structure.
Before investing in artificial intelligence, leadership needs to define which change deserves to be explored, why it matters and what evidence would help decide whether to continue.
The choice of tool makes sense within that direction.
What possibility are we not considering yet?
When we think about AI from the current routine, it is natural to look for activities to speed up: producing a presentation, summarizing a meeting, preparing a proposal.
Those are possible uses. But whoever runs a business also needs to make room for questions about how the company itself works.
Why does a proposal have to go through so many steps? What information reaches leadership only after the opportunity has passed? What kind of follow-up would we like to offer customers, but find hard to sustain today?
These questions widen the conversation. The investigation stops being limited to the time spent on an activity and starts to include the value it should produce.
For a founder, a director or a board member, this may mean revisiting an assumption that has guided decisions for years: something that used to require a certain structure might now be organized another way.
That possibility needs to be examined. It is not yet a promise of results.
Question: understand what really needs to change
Consider an illustrative situation: a company wants to use AI to prepare sales proposals faster.
Before looking for a solution, leadership asks the team to examine a few recent proposals. Where were the delays? At which steps did questions come up? What had to be redone?
The investigation may show that writing takes up only part of the time. The team may receive incomplete information, struggle to define the scope or depend on unclear approvals.
In that scenario, speeding up the writing would solve part of the problem. The rest would keep producing waiting and rework.
Questioning means making visible what we are trying to solve — including when the original request needs to change.
A useful question for this conversation is:
If this task were done in a few minutes, what would still stand in the way of the result we want?
Redesign: imagine another way of working
With the problem clearer, the team can explore other possibilities.
In the proposal example, could the information be organized better before drafting? Could gaps be identified while the conversation with the client is still happening? Could alternatives be prepared so the sales team can evaluate different paths?
AI could take part in some of these steps, depending on the data available, how well the technology fits and the review required.
Leadership does not need to walk into the meeting knowing how to configure this process. It needs to be able to state a direction clear enough to guide the investigation.
Instead of asking only for a list of tools, it can ask:
Show us how we could improve the handoff between understanding the client and presenting a proposal. Where could AI contribute? What should stay under the team's judgment?
That request makes room for the knowledge of the people who do the work and lets you compare possibilities before choosing a solution.
Rebuild: turn the possibility into a concrete experiment
An idea has to meet the company's real conditions.
In the example, the team could test a new way of preparing proposals on a limited set of cases, with approved information and review before anything is sent.
Before the test, the team would need to agree on what to watch: clarity of scope, quality of the proposal, total preparation time and number of corrections, for example.
It is also important to decide who follows the experiment and who decides the next step. A test can lead to expansion, to adjustments or to the conclusion that the approach is not worth continuing.
Rebuilding is that passage from imagining to experimenting. The experiment gives leadership information to decide on firmer ground.
What to ask the team before approving an AI investment
A proposal to use artificial intelligence is easier to evaluate when it answers five questions:
- What change do we want to produce?
Describe what would improve for the business, for the customer or for the people who work there. - What do we know about the current situation?
Present examples, difficulties and available information, separating observations from assumptions. - Which alternatives will we explore?
Consider changes to the process and possible contributions from AI. - How will we evaluate the experiment?
Define quality criteria, limits of use and situations that require human review. - What decision will we make at the end?
Agree on which evidence would justify continuing, adjusting or stopping.
These questions help turn interest in AI into an investigation with a purpose.
A leader's view also needs room to change
Building a business means recognizing possibilities before you have all the answers. Leading also requires judging which of them deserve attention, resources and responsibility.
That is where I place AI Thinking: Question, Redesign and Rebuild as moves to rethink work and explore what can change with artificial intelligence.
A keynote or a training session can open that space: present a concrete possibility, examine its implications and help shape the questions leadership will take to the team.
The starting point can be a question for your next meeting:
What would we like to make possible in the business, but are still evaluating with the limits of the past?
Dani Hayoshi is the author of AI Thinking. Her work connects artificial intelligence, education and entrepreneurship. About Dani.