A different question changes the conversation
When a new technology appears, it is natural to ask which tasks it can perform. Leadership’s question is broader: what customers need, what constrains the operation and which combination of people, processes and systems produces a better result. Sometimes the answer includes AI; sometimes it is simplification.
The OECD’s study of SMEs identifies barriers involving fit with work, data, legal issues and skills. Access to licenses alone does not resolve them.
From individual users to shared work
Individual learning matters, but an organization needs agreements: what information may be used, how an answer is checked, where decisions are recorded and who receives an escalation. People who know the process should participate in its redesign.
Explain why work is changing, which decisions remain human and how quality will be assessed. Training should connect to real tasks and opportunities to practice, not only a general presentation.
Build trust through verifiable boundaries
A model can produce a convincing but incorrect answer. Define acceptable evidence, review proportional to risk and information protection. NIST offers a voluntary profile for managing generative AI risk across its lifecycle.
In a hypothetical proposal-writing test, a specialist could verify figures and terms before they are shared. Responsibilities should be clear even if initial preparation is automated.
A management routine for learning
Review which decisions improved, which exceptions increased, what skills are missing and what customers or colleagues report. Acknowledge negative results and change a hypothesis when evidence warrants it.
The goal is not adoption for its own sake. It is building judgment and a way of working that can capture opportunities responsibly. Jevons can support diagnosis, design and a change roadmap.
Frequently asked questions
How can adoption be encouraged without mandating a tool?
Involve people in real problems, explain the expected benefit, allow questions and measure quality as well as use. An adoption rule cannot substitute for training or work design.
Which capabilities should be developed?
Beyond tool use, important capabilities include problem framing, evidence evaluation, process knowledge and recognizing when to escalate a risk or error.
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