AI Upskilling: Building an Organization That Learns
How individual development and organizational learning can reinforce each other to create lasting value from AI.
When we talk about upskilling for AI, we often start with the tools. Which platforms should people learn? How can they write better prompts? Which tasks could they complete faster?
These are useful questions. People need practical experience with AI and an understanding of its capabilities and limitations. But our ambition should extend further.
Are we developing people who can work with AI, and an organization that can learn from what they discover?
An individual may find a better way to solve a problem, recognize a recurring error or discover an entirely new opportunity. What happens to that insight? Does it stay with them? Get shared in a meeting? Or change how the organization works?
That connection between individual learning and organizational learning is where I see an opportunity for AI upskilling to create lasting value.
Develop the thinking through the work
AWS’s Reimagine: Turning AI into Value identifies five durable skills: learning and unlearning agility, problem framing, judgment, systems thinking and curiosity. It emphasizes developing these capabilities through practical work, feedback and progressively more demanding challenges. (AWS Reimagine, Chapter 4)
Each changes what effective AI learning looks like.
Problem framing means getting clear about the outcome before asking AI for a solution. Judgment means evaluating what comes back, identifying what is missing and deciding whether there is enough evidence to act. Systems thinking means considering what happens beyond the task immediately in front of us.
Learning and unlearning require us to examine familiar approaches: why did we develop them, which assumptions still hold and what deserves to change? Curiosity opens up the possibilities we might otherwise miss.
These capabilities need practice. Give someone a real problem, let them work through it with AI and ask them to explain their decisions. What did they accept? What did they challenge? What changed their mind?
Doing remains essential to learning. The design of the work determines what people have an opportunity to learn.
Connect individual learning to organizational learning
AI upskilling becomes more valuable when individual learning strengthens organizational learning, and the organization makes that learning available for people to build on.
People bring discoveries, questions and feedback from their work. The organization tests those insights, improves shared practices and evaluates the results. In return, people gain access to shared knowledge, better learning conditions and new challenges.
People learn through work. Organizations learn when those insights change how work happens.
Consider a hypothetical example: a team using AI to prepare customer proposals.
One person notices that the drafts repeatedly overlook a delivery constraint. They learn to supply better context and check the result more carefully. Their individual capability improves.
Then they bring the issue to the team. Colleagues compare examples and discover that the constraint is poorly documented and interpreted differently across functions.
The organization now has something to learn.
The team clarifies the requirement, updates the shared guidance and assigns responsibility for keeping it current. Future proposals start with better context. Other colleagues can build on that improvement, and their experience helps reveal whether it works or needs further revision.
One person’s discovery becomes a shared improvement, then gets tested through use.
The essential step is applying what was learned. A growing collection of documents or a well-attended show-and-tell can support the process. The loop closes when insights change decisions, practices or systems, and people examine the effects.
This is the learning organization in practice
The connection has an established foundation in Peter Senge’s work on the learning organization. His five disciplines bring together personal mastery, mental models, team learning, shared vision and systems thinking. They connect individual development with the ability to question assumptions, think collectively and understand the wider system. Society for Organizational Learning
David Garvin similarly emphasizes that learning organizations act on new knowledge by changing their behavior. Harvard Business Review
AI upskilling offers a practical opportunity to put these ideas to work.
People need space to experiment and reflect. Teams need ways to examine what individuals discover. Leaders need to respond when those discoveries reveal an outdated process, an unclear decision right or an incentive that discourages a better approach.
The organization also feeds learning back to the individual. Better guidance, access to colleagues’ experience, clearer boundaries and more useful feedback give people a stronger foundation for their next challenge.
Both sides need attention. Asking people to learn continuously has limited effect if the organization cannot change in response.
How learning can become a flywheel of value
Reimagine describes several mechanisms that support this connection: peer learning, workflows that become reusable team assets, and shared context that allows new projects to benefit from accumulated organizational knowledge. (AWS Reimagine, Chapters 3 and 9)
I see the possibility of a flywheel here.
People build capability through practice. Their insights improve shared ways of working. Those improvements help others work and learn more effectively. Better outcomes create a reason to continue investing in the learning process.
An author’s synthesis informed by Reimagine and learning-organization principles. Value can compound when lessons are applied, outcomes improve and learning continues.
That is a possibility to design for and test. Sharing an insight does not guarantee that it is correct or useful everywhere. A practice that helps one team may create problems elsewhere. The organization needs to evaluate what it learns, understand where it applies and revisit it as conditions change.
The flywheel depends on the quality of those feedback loops.
Make the loop part of everyday work
For leaders, this changes the design of an AI upskilling program. Alongside teaching people how to use tools, we need to create a route from individual discovery to organizational action.
A practical starting point is one team working on one meaningful problem:
Define the outcome. Agree on what should improve for the customer, employee or business, and how the team will assess it.
Practice on real work. Give people appropriate tools, boundaries and time to experiment.
Discuss the reasoning. Review what worked, what failed, what was verified and what remains uncertain.
Apply the lesson. Decide whether to change a shared practice, workflow or assumption, with someone responsible for the change.
Check the effect. Establish whether the change improved the outcome and what the team should try next.
Leaders’ responses matter throughout. If people are rewarded only for producing more, reflection can feel like a distraction. If surfacing a mistake carries a penalty, useful information may never reach the people who need it.
We also need to preserve opportunities for less experienced colleagues to develop judgment. Reviewing AI output with an experienced colleague, debating alternatives and working through realistic cases can create deliberate opportunities to learn. Reimagine highlights this as an unresolved challenge as AI changes the tasks through which junior employees traditionally gained experience. (AWS Reimagine, Chapter 4)
Connect learning to efficiency and opportunity
This approach supports both AI for efficiency and AI for opportunity.
A team may begin by reducing rework or improving the quality of a familiar task. As its understanding grows, it can ask a broader question:
What else could and should we be doing?
Could we address a customer need we previously lacked the capacity to serve? Explore an option we had dismissed as too difficult? Redesign the entire process now that an old constraint has changed?
Those possibilities become the next learning challenge.
For an independent professional or small business, the same principle applies: reflect on experience, update how you work, learn with peers and test whether the changes help. Within a larger organization, the additional responsibility is to make learning travel across teams and influence the systems around them.
The ambition for AI upskilling should be that people become more capable, their discoveries improve the organization, and the organization helps them keep developing.
When one person learns something valuable, how does it make the next person, and the organization, more capable?
