Understand the situation.
Look closely at the work, the pressures around it and what needs to improve before settling on an answer.
How we approach the work
Begin with the work. Leave the organisation stronger.
We are interested in change that improves the immediate result and how the organisation works afterwards.
Develop the response.
Bring together evidence, experience and different perspectives to design something that works in the organisation’s actual context.
Make the change work.
Put the response into practice, learn from what happens and strengthen the organisation’s ability to continue without us.
Human–AI Work
Better performance and stronger human capability belong together.
AI can increase speed, range and analytical power. The more important question is what happens to the quality of the work and to the people doing it.
We consider how human and artificial intelligence contribute, where judgment and responsibility should remain clear, and how people continue developing the expertise to question and improve what AI produces.
Work
What should people and AI each contribute?
Team
How should colleagues collaborate, judge and learn?
Organisation
What must leadership, roles and systems support?
Organisational Decision Systems
A decision has a life beyond the meeting.
Important decisions are shaped before they are approved and tested afterwards. We look at whether the choice has been framed and tested well, whether approval creates real commitment and what happens when the decision meets reality.
Form
What decision are we making, and what would justify it?
Define the choice, examine the evidence and uncertainty, and compare credible alternatives.
Commit
What changes because we have decided?
Make authority, resources, priorities and responsibilities explicit.
Hold
How will we keep it on course and know when to change it?
Review results and test the assumptions, address exceptions and decide explicitly whether to continue, adapt or reopen.
Across the lifecycle, we look at the information, authority, behaviour, resources and learning that make a decision real in the organisation.
The Hidden Life of Decisions develops this perspective in greater depth.
Our point of view
The useful question is not only what AI can do. It is what the work should become.
AI can widen the possibilities, improve an immediate result and help people contribute beyond their usual specialism. But better output alone does not guarantee better judgment, clearer responsibility or stronger expertise. Those outcomes depend on how the work around AI is designed.
Separate performance from understanding.
AI-assisted work can improve before people’s understanding improves with it. Management needs to look at the quality of the result and at whether colleagues can assess, explain and improve it.
Fit the relationship to the work.
AI should not play the same role everywhere. The right arrangement depends on the task, the reliability required, the consequences of error and where human knowledge adds something different.
Design capability and accountability together.
Roles, review, authority and opportunities to learn determine whether AI becomes part of better work or simply another layer of output for people to check.
Together, these findings move the management question beyond adoption. The design task is to improve the work now without weakening the judgment, ownership and capability the organisation will need next.
Explore the Human–AI Work SystemCurrent thinking
New capability changes the work around it.
AI can improve what organisations produce. It also changes what people need to understand, how judgment is exercised and how organisations turn decisions into action.
Human–AI Work
What must people still understand when AI does more of the work?
A better output can obscure what its author has understood. How can organisations develop the expertise to assess AI-assisted work?
Read article
Human–AI Work
Using AI is not the same as working well with it.
Look beyond adoption and time saved to the quality of the work and the demands it places on colleagues.
Read article
Decision-making
Why an agreed decision may still change nothing.
An approved priority may leave existing commitments untouched. What must management resolve before the business can act?
Read articleResearch watch
What we are learning from the field.
Human–AI work is developing quickly. We follow new evidence across cognitive science, human factors, organisational learning and management practice to test our thinking and identify the questions leaders will need to address next.
Expertise and learning
Performance is not the same as learning.
AI improved the work of patent lawyers, but the development of expertise differed by prior experience. The management question is therefore wider than productivity: what are people learning while AI helps them perform?
Field experiment · David Autor and colleagues, 2026
View the research
Collaboration and expertise
AI can change how expertise moves through a team.
Research with P&G professionals found that AI helped individuals cross functional boundaries and changed the experience of collaborative work. This raises practical questions about team composition, review and shared judgment.
Field experiment · Fabrizio Dell’Acqua and colleagues, 2026
View the research
Organisational response
Faster analysis increases the value of acting and reallocating.
Better information creates little value when priorities, resources and commitments remain fixed. As AI accelerates analysis, organisations also need to become better at changing course.
Management survey · McKinsey, 2026
View the research
Explore the work
