People & adoption

What needs to change in the work?

Explore skills, responsibilities, workflows, and the decisions people need to own.

A PLACE TO BEGIN

AI Made Tools Easy; Judgment Got Hard: How Strong Teams Operate Differently

An article on the judgment and trade-offs people need to own when AI can generate the options.

Read the article →

What does AI adoption mean in day-to-day work?

AI adoption is the change in how people carry out work, use information, and make decisions with an AI-enabled capability. Giving a team access to a tool is one step. The business outcome depends on whether the tool fits the work, whether people can judge its output, and whether responsibilities and support are clear.

The people who know the business best should help define the problem, test the approach, and decide what needs to change. Their knowledge of exceptions, customer expectations, and operational constraints is part of the implementation, rather than something to consult only after deployment.

Connect AI output to a decision someone can make

Models do not repair unclear ownership, slow escalation, or broken workflows. They can surface those weaknesses at greater volume. When a signal triggers a search for the person authorized to act, a technical improvement can become additional coordination work.

Describe the workflow from input to action. Who reviews the output? Who accepts responsibility for the decision? What can they do without another approval? What happens when the model is wrong, the information is incomplete, or no one is available? The operating-model readiness question and the costs of unclear ownership examine this gap between detection and action.

Build the skills the changed work requires

As AI takes on parts of information gathering or drafting, people may spend more time framing problems, checking evidence, handling exceptions, and weighing trade-offs. Training should reflect those responsibilities. It should include examples of good and poor outputs, practice with realistic exceptions, and a clear way to question or override a recommendation.

How judgment changes when tools become easier explores these decisions. The reallocation planning worksheet helps make the discussion concrete: which tasks change, what time is released, and what useful work should receive that time?

An illustrative example: a new review assistant

A team may test an assistant that prioritizes cases for review. Adoption requires more than accepting its ranking. Reviewers need to understand what the ranking can and cannot tell them, how to handle an urgent case it misses, and how their feedback changes the process. Managers need to account for review capacity and watch whether pressure to process cases undermines quality. These are implementation questions, rather than claims about a particular deployment.

Treat adoption as a learning process

Start with a bounded workflow, observe how people actually use it, and review errors and workarounds with the team. Make it possible to pause or revert the change. Track quality, rework, workload, and customer experience alongside usage. A high login rate does not establish business value; a low usage rate may reveal a workflow problem worth investigating. Connect those observations to the outcome and cost of the change.

Articles on people & adoption

Search this topic →
Browse all 55 articles in this topic →

Explore another question.

OpportunitiesWhere could AI make a difference? →Investment decisionsWhat deserves investment? →Outcomes & evidenceHow do we judge the result? →