Part 4 of The AI Skills Paradox Series
Chief Learning Officer "June" — Follow Her Story
Over the course of this series, June's understanding of artificial intelligence had changed considerably. She had started by asking how her organization could use AI to create learning faster, personalize experiences, and increase the productivity of her team. Those opportunities remained important, but they increasingly felt like only one part of a much larger transformation.
June had learned that abundant content doesn't automatically create capability. She had seen how removing too much friction can eliminate the productive struggle through which people develop skills and judgment. She had also discovered that as AI makes content easier to produce, understanding the business, the learner, and the behavior that needs to change becomes more valuable.
Together, those insights led her to a more fundamental question about the purpose of Learning & Development itself. If AI can increasingly create courses, scenarios, assessments, communications, and even feedback, what is the work that remains uniquely important for L&D?
The answer required June to separate what her organization creates from what it is ultimately responsible for developing. Courses, content, technology, programs, and learning experiences are important outputs, but none of them are the actual business outcome.
The real work is developing human capability.
Learning Is Not the Business Outcome
For decades, the language of L&D has naturally centered on learning. Organizations build learning strategies, implement learning platforms, create learning experiences, measure learning engagement, and employ learning professionals. Those terms accurately describe much of the work the function performs.
Business leaders, however, rarely wake up wishing their organization had more learning.
They need salespeople who can navigate more complex customer conversations. They need managers who can coach employees effectively. They need leaders capable of making difficult decisions during uncertainty. They need teams that can adopt new technology, work differently, solve emerging problems, and respond quickly when business conditions change.
Learning matters because it helps create those outcomes.
That distinction becomes increasingly important as AI lowers the cost of producing learning. If organizations define L&D's value primarily through courses, content, and programs, then increasingly capable AI can appear to threaten the function's relevance. If L&D's purpose is instead defined by its ability to help people become more capable at things that matter to the business, AI becomes another powerful tool for accomplishing that mission.
Unboxed President Dave Romero captures the distinction directly in his memo to employees: clients aren't ultimately trusting learning partners simply to produce training. They are trusting them to help their people "become better at something that matters to their business." The Work We Are Really In
That is a fundamentally different way to define the work.
Capability Begins With What People Need to Do Differently
Once June accepted capability as the objective, the starting point for L&D changed as well.
Instead of beginning with a learning request, her team began with business performance. They worked backward from the outcomes the organization needed and identified the behaviors and capabilities required to produce them.
If the business wanted stronger sales performance, the conversation didn't immediately begin with sales training. It began by understanding how customer conversations needed to change. If the organization wanted stronger managers, the team identified what effective management actually looked like in practice. If leaders needed employees to adopt AI effectively, the question wasn't simply whether employees knew how to use the tools, but whether they could apply them with appropriate judgment in the context of their work.
This creates a relatively simple sequence for L&D:
- Business strategy identifies what the organization needs to accomplish.
- Performance requirements define what people need to do to execute that strategy.
- Critical capabilities identify the skills and behaviors required for successful performance.
- Development experiences help employees build those capabilities.
- Measurement determines whether capability and performance are actually improving.
This is very different from beginning with a request for a course.
Dave's memo proposes a similar diagnostic approach by asking what people are doing today, what needs to be different tomorrow, why the desired behavior isn't already happening, and what an individual needs to experience in order to become more capable. The Work We Are Really In
For June, those questions were becoming the foundation of a new L&D operating model.
Capability Development Requires More Than Learning
The shift also changed what happened after a learning experience was created.
Traditional learning strategies often concentrate resources around the moment of instruction. Employees attend a program, complete a course, participate in a workshop, or consume digital content. Once the learning event concludes, responsibility for applying the material often shifts almost entirely to the employee and their manager.
Capability development requires a broader system.
Employees may first need foundational knowledge, but they also need opportunities to apply that knowledge in increasingly realistic situations. They need to practice, receive feedback, observe others, reflect on their performance, try different approaches, and receive coaching as they move from understanding a concept to performing consistently.
That creates a continuous development cycle:
- Learn the concepts, frameworks, and knowledge required for performance.
- Connect with peers, experts, mentors, and managers who provide context and perspective.
- Practice the behaviors in realistic, lower-risk situations.
- Receive feedback that identifies strengths and development opportunities.
- Coach against observable behaviors and specific performance goals.
- Apply the capability in the flow of real work.
- Measure whether skill, behavior, and performance are improving.
- Adapt development as individual and business needs change.
No individual component is sufficient by itself. A course can establish knowledge without changing behavior. Practice can become repetitive without meaningful feedback. Coaching can become subjective without clear performance expectations. Measurement can become meaningless if the organization hasn't defined which capabilities matter.
The value comes from connecting these experiences around a common definition of capability.
AI Changes the Means, Not the Mission
Throughout her journey, June had never become less enthusiastic about artificial intelligence. She had become more precise about what she wanted AI to accomplish.
AI could remove unnecessary production work, accelerate research, organize knowledge, generate first drafts, create scenarios, personalize reinforcement, support practice, analyze conversations, and surface patterns that would be difficult for humans to identify manually. These capabilities could allow her organization to operate at a scale that would have been impossible only a few years earlier.
But none of those capabilities changed the underlying mission.
Dave makes this distinction particularly clear in his memo. He encourages Unboxed to use AI aggressively, automate work that should be automated, and create experiences that previously couldn't be created at scale. At the same time, he argues that determining what should be automated requires judgment, curiosity, empathy, and an understanding of the people the experience is designed to develop. The Work We Are Really In
June began thinking about AI in exactly those terms.
The goal wasn't to insert AI into every stage of learning. The goal was to determine where AI could make capability development more effective, more personalized, more scalable, or more measurable without removing the experiences people needed in order to grow.
Sometimes AI should provide the answer.
Sometimes it should ask the question.
Sometimes it should create the scenario.
Sometimes it should play the customer, employee, or manager inside that scenario.
Sometimes it should provide feedback.
And sometimes the most intentional design decision may be requiring the learner to wrestle with the problem before AI provides any assistance at all.
The technology is extraordinarily powerful. The learning strategy still determines how that power should be used.
Practice Becomes More Important as Information Becomes Easier to Access
AI also forced June to reconsider the relationship between knowledge and skill.
Employees increasingly have access to extraordinary amounts of information at the moment they need it. They can ask AI to explain a concept, summarize a document, prepare for a conversation, analyze information, or recommend possible actions. In many situations, memorizing information that can be retrieved instantly becomes less important.
That doesn't make expertise irrelevant.
It changes what expertise looks like.
A manager can ask AI for five questions to use during a coaching conversation, but still needs to recognize which question is appropriate in the moment. A salesperson can receive a detailed account briefing before a customer meeting, but still needs to listen, build trust, adapt, and exercise judgment during the conversation. A leader can use AI to analyze several strategic options, but still owns the decision and its consequences.
Those capabilities develop through experience.
This makes practice increasingly important in the AI era, not less. When information becomes readily available, organizations can devote more developmental energy to helping people apply information under realistic conditions. AI-powered simulations, roleplay, scenario-based learning, and repeated practice can expose employees to situations they might otherwise encounter only occasionally in the workplace.
The objective is not to prove that people can perform without technology. It is to ensure they can perform effectively with technology while retaining the human judgment required to use it well.
Managers Become Part of the Learning System
June's new model also made it impossible to treat managers as separate from L&D.
Employees spend only a small portion of their working lives inside formal learning experiences. Their managers influence performance continuously through expectations, observation, feedback, prioritization, and coaching.
If the learning organization teaches one set of behaviors while managers reinforce another, the manager usually wins.
Capability development therefore requires managers to understand which skills the organization is trying to build and how those capabilities should appear in everyday performance. They need tools for observing behavior, frameworks for providing feedback, and enough visibility into employee development to know where coaching can have the greatest impact.
AI can strengthen this relationship as well. Practice data can surface development opportunities before a coaching conversation. Conversation analysis can identify recurring behaviors. Skill data can help managers distinguish between isolated mistakes and patterns that require focused development.
The manager doesn't disappear from the learning system as AI becomes more sophisticated. The manager becomes increasingly important because someone still needs to connect development with the context of actual work.
Capability is ultimately developed through the combination of learning, experience, technology, and human reinforcement.
Measurement Must Move From Activity to Capability
Perhaps the most significant implication of June's new perspective involved measurement.
If L&D's purpose is to deliver learning, traditional metrics make sense. Course completions, participation, satisfaction, learning hours, assessment scores, and content engagement provide useful information about whether employees are interacting with learning programs.
If the purpose is to develop capability, those measures become incomplete.
June still wanted to know whether employees participated in learning, but she increasingly wanted answers to different questions:
- Can employees demonstrate the skill?
- Is their performance improving with practice?
- Which capabilities are developing most quickly?
- Where do significant skill gaps remain?
- Are managers reinforcing the expected behaviors?
- Can employees transfer what they practiced into real work?
- Which capabilities are associated with stronger business performance?
- How should development priorities change as the business evolves?
These questions move L&D from learning analytics toward Skill Intelligence.
Instead of simply reporting what employees consumed, organizations begin building visibility into what their people can actually do. Practice, coaching, assessments, observations, performance data, and AI-generated insights can contribute to a more complete picture of capability development.
This doesn't make measurement easy. Connecting human capability to business outcomes remains complex, and L&D should resist overstating causation simply because more data becomes available.
But the direction matters.
The goal is to measure closer to the thing the organization is actually trying to develop.
The Future L&D Organization Is a Capability Engine
By the end of this journey, June's view of her organization had expanded considerably.
Her team still created learning. They still designed courses, built programs, implemented technology, supported managers, and managed learning operations. Those activities weren't disappearing, and many would become significantly more powerful with AI.
What changed was how June understood their purpose.
The learning organization existed to help translate business strategy into human capability. Its role was to understand where performance needed to change, identify the capabilities required, create meaningful development experiences, provide opportunities for practice, enable coaching, and measure whether people were actually becoming better at the work that mattered.
That perspective also changed how June thought about AI. She no longer evaluated the technology primarily by how much time it saved or how much content it could produce. She evaluated it by whether it helped the organization develop capability more effectively.
In his memo, Dave draws a clear distinction between the outputs of learning work and its ultimate purpose: "Content, Courses, Software, Implementations, Timelines, and Strategy is what we create. Human capability is what we develop." The Work We Are Really In
For June, that became more than a statement about instructional design.
It became a definition of the modern L&D function.
What the AI Skills Paradox Really Means
Across this series, June encountered what initially appeared to be several separate challenges.
AI made content dramatically easier to produce, yet more content didn't necessarily create greater capability. Technology made learning increasingly frictionless, yet some forms of effort were essential to development. AI generated increasingly sophisticated experiences, yet understanding which experiences mattered required deeper context and judgment.
Those challenges were all expressions of the same underlying paradox.
The easier AI makes it to produce the outputs associated with learning, the more important it becomes to understand the human process through which capability actually develops.
That doesn't require L&D to resist AI.
It requires L&D to use AI for a more ambitious purpose.
The opportunity isn't simply to create the same learning faster. It is to redesign development around what technology now makes possible while protecting the practice, judgment, relationships, feedback, and experience through which people become capable.
That means removing friction when friction serves no purpose and preserving productive struggle when effort contributes to growth. It means using AI to create content while investing human attention in context. It means automating production while becoming more rigorous about diagnosis. And it means measuring success through capability and performance rather than the volume of learning produced.
The technology may change continuously.
The responsibility of L&D remains remarkably consistent.
Help people become better at something that matters.
Key Takeaways
Artificial intelligence is transforming nearly every aspect of enterprise learning. It can dramatically accelerate content production, make individualized practice scalable, provide immediate feedback, support managers, analyze performance, and create development experiences that would previously have been impossible to deliver across large organizations.
The strategic opportunity, however, is larger than efficiency. As content becomes abundant and information becomes easier to access, L&D's value increasingly shifts toward diagnosing performance problems, identifying the capabilities that matter, designing meaningful experiences, preserving productive struggle, enabling managers to reinforce development, and measuring whether people can actually perform.
The organizations that navigate the AI Skills Paradox successfully will not choose between human expertise and artificial intelligence. They will build development systems in which each makes the other more effective. AI will provide scale, speed, personalization, insight, and new opportunities for practice, while human judgment provides context, intention, empathy, business understanding, and meaning.
Dave's memo ultimately describes the responsibility in straightforward terms: clients trust learning partners with something more important than producing training. They trust them to help their people become better at something that matters to the business. The Work We Are Really In
That is the opportunity AI creates for L&D. Not to become a faster content factory, but to become a more effective engine for human capability.
And that is the work we're really in.
