The question is a business one
A company can have licenses, pilots and vendors without having a strategy. Leadership’s decision is not how many tools to buy, but which objectives deserve resources, which constraints limit performance and which risks the business can accept. AI is one possible capability in that conversation.
The OECD notes that adoption can require changes in culture and practice alongside suitable skills and data. A tool does not resolve those dependencies by itself. Priorities should connect to a business measure and an accountable owner, not just an impressive demonstration.
Five decisions that organize the agenda
First, define the outcome: profitable growth, quality, capacity or risk management. Second, choose processes or decisions where the problem matters and a baseline exists. Third, compare redesign, conventional rules, training and AI. Fourth, establish who decides, reviews and owns the result. Fifth, agree what evidence would justify scaling, correcting or stopping.
These choices do not require solving the entire transformation at once. A small portfolio, differentiated by value, uncertainty and data dependencies, lets leaders learn without confusing activity with progress.
From portfolio to governance
A leadership team can review a concise monthly dashboard: value hypothesis, full cost, quality, adoption, open risks and next decision. NIST’s generative AI profile provides a voluntary framework for considering risks across the lifecycle; it needs to be adapted to context and is not automatic certification.
For example, a hypothetical distributor might prioritize forecast quality and quotation turnaround before a broad corporate platform. Each initiative needs a business owner and a route for escalating exceptions to someone with authority.
A decision to make this week
Select one decision that currently commits money or capacity. Identify its owner, missing information and the result that would justify action. Jevons can help frame alternatives, decision criteria and a roadmap; design, coordination and implementation responsibilities are agreed for each engagement.
A useful strategy also says what will not be done yet. Waiting, improving data or simplifying a process may be the right decision.
Frequently asked questions
Who should lead AI strategy?
An executive sponsor should connect priorities to business outcomes, while operations, technology, data, risk and people teams contribute to the decision. Accountability does not disappear when a vendor is engaged.
Should the highest-impact case always come first?
Not necessarily. Compare potential value with data quality, learning time and risk. A meaningful but bounded opportunity can produce evidence before committing to a broader transformation.
Sources and further reading
Let’s discuss your next decision.
Tell us what is at stake, what remains uncertain and where a clearer strategic direction would help.
Start a conversation