The AI value framework applied to learning and development

Why work an example

The three paths are easy to agree with and hard to apply. Teams sort their own candidates onto the wrong path, usually by filing a productivity gain as a disruption because it feels ambitious. Working one function end to end makes the boundaries concrete.

Boost productivity: the same work, faster

• Content development that ran in weeks running in hours, including variants of one module for different audiences.
• Translation and localization at a cost that makes a multi-market rollout a decision rather than a project.
• Practice scenarios and assessments drafted from the material that already exists.
• Custom visuals and voiceover produced without a studio booking or a designer in the queue.

This is where most functions start, and it is a real gain. It is also where most stop, which is the failure mode the framework is built to catch: tools rolled out, workflows unchanged.

Create value: something the function could not offer before

• Learning advisors that answer a person mid-task instead of sending them to a course catalogue.
• Role-play and simulation that responds to what the learner actually said, which is the part a static scenario cannot do.
• Paths that adapt to progress rather than to a job title.
• Support delivered inside the tools where the work happens, at the moment the work is happening.
• Skill assessment that names a specific gap and the next step against it.

These are new offerings rather than faster versions of old ones. The test is whether a learner could have got the same thing before and simply waited longer. If they could, it belongs on the first path.

Disrupt models: who does the work changes

• A small team producing and maintaining content at a scale that used to require a department or a vendor.
• Instructional design compressed from months to days, with revision built into the cycle rather than scheduled as a project.
• Expert knowledge captured and structured once, reducing the dependency on an expert being available.
• Capabilities that used to require an enterprise platform assembled from ordinary tools.

This path is uncomfortable because it puts the function’s own shape in scope, including which parts are bought in. It is also where the durable advantage sits, and it is the reason the three paths are scored separately rather than ranked against each other.

Applying this to your own function

1. 1
   List the candidates without sorting them

   Collect what people have already tried, asked for, or seen elsewhere. Sorting during collection loses the ones that sound too small.

2. 2
   Place each one on a path

   Ask what the person on the receiving end could have got before. Same thing faster is the first path, something previously unavailable is the second, a change in who produces it is the third.

3. 3
   Compare within a path, not across

   A productivity gain and a business model change do not belong on one ranked list, and comparing them is how a portfolio collapses into a pile of quick wins.

4. 4
   Name an owner per path

   The first path belongs to functional leaders, the second to product and function together, the third to the executive team. A path with no owner at the right level does not move.

How this is delivered

Use-case discovery runs this exercise on the client’s own candidates and leaves a prioritized portfolio with the reasons for what did not make the cut.