Leadership Insights
Adoption is a leadership question before it is a technology question
Teaching you can act on without buying anything. Where research is quoted it is named, dated and linked, so you can read the original rather than take our word for it.
The five resources
Successful people and organisations learn to make better use of five resources: time, people, money, knowledge and technology. Agentic AI creates new opportunities to bring those resources together.
- Time
- Notice where repetitive work consumes attention that belongs on judgement and relationships.
- People
- Skills, judgement, confidence and clear responsibility. Capability is developed, not assumed.
- Money
- Weigh the value of a change against what it costs to implement and to run.
- Knowledge
- Capture the expertise your organisation relies on, and keep it current and easy to apply.
- Technology
- Connect tools to the way work actually happens, rather than to the way a product demonstrates.
Knowledge is the one most often left informal. An agentic system applies written knowledge quickly, so the clearer your organisation is about how it does things, the more useful that system becomes.
A first move that costs nothing: take the ten questions your team asks each other most often this month, write the agreed answer to each, and decide who keeps them current.
What the published research says
Two pieces of published work are worth a leader's time. Both are linked in full. Neither is a MAP result and we do not present them as one.
MIT NANDA, The GenAI Divide: State of AI in Business 2025
Published in July 2025 by Project NANDA, a project of the MIT Media Lab, and written by Aditya Challapally, Chris Pease, Ramesh Raskar and Pradyumna Chari. The authors describe it as preliminary findings, drawn from a review of more than three hundred publicly disclosed initiatives, interviews with fifty two organisations and survey responses from one hundred and fifty three senior leaders.
Within that sample, most initiatives showed no measurable financial return, and the authors attribute the difference to approach rather than to model quality or regulation. The widely repeated headline figure describes those initiatives, not organisations in general, and the paper has been publicly disputed, so we quote no percentage here. The leadership lesson stands on its own: a pilot with no owner, no training and no measure of success rarely survives contact with the working day.
Read the reportMcKinsey, The state of AI
McKinsey's ongoing survey work makes a similar point from the other direction: the organisations reporting real effect tend to be the ones that redesigned a workflow and made somebody accountable for it, rather than the ones that simply bought more capability. Read the current edition and the sample it describes before applying any figure to your own organisation.
Read the reportActivity is not progress
A team's record can look excellent. Tasks completed, notes written, a full diary, nothing overdue. Then you ask which relationships are further along than they were a month ago, and the answer is almost none. Where automation writes the record, a convincing account of activity can exist with little movement behind it.
At your next review, do not ask what was done. Ask what moved, and what caused it.
Purposeful, the podcast
Listen to Purposeful mini podcasts for short reflections on leadership, mindset and making better use of the resources available to you.
