Use this whenYou’re setting up an AI program and deciding where your team’s time goes.

Why it exists

Most AI programs blur two very different jobs. Training gets judged on ROI it was never designed to produce, and gets cut. Or a central team tries to deliver every project itself and becomes the bottleneck. Separating the two lets you staff, fund, and measure each one properly.

The framework

EnablementIs the organization able to start?Enablement as recruitingMeasured by reach and the talent it surfaces
ExecutionWhat tangible return did this deliver?Measured by tangible return
  1. Enablement

    Is the organization able to start?

    Training, AI councils, champions programs, and building kits that get people using AI in their own work. It gets the organization off the ground. Measure it by participation, by what teams build for themselves, and by the people it surfaces.

  2. Enablement as recruiting

    Who has the aptitude to pick this up?

    Every session is also a search. The people who lean in, come back with questions, and start building are your future champions and builders. Enablement is how you find them inside the organization, without a job posting.

  3. Execution

    What tangible return did this deliver?

    Redesigning real workflows, building agents, and changing how the business runs: the digital transformation. Measure it in spending avoided, capacity released, and work removed, with the checking and upkeep counted honestly.

How to use it

  1. Sort your AI work into enablement and execution. Most teams find they’ve been measuring both the same way.
  2. Set different measures: reach and talent for enablement, tangible return for execution.
  3. Use enablement to spot people with the aptitude, then bring them into execution work as champions or builders.
  4. Pick execution work with the data, process, people lenses, and test it with a field-test worksheet before scaling.
  5. Feed execution wins back into enablement as real examples. They teach better than any demo.

Two ways this goes wrong

  • Training on trial: enablement gets judged on ROI it was never designed to produce, so it gets cut before it finds anyone.
  • The central bottleneck: one team tries to deliver every project itself, and every improvement waits in its queue.

How the split shifts in the first year

  • Early on, a typical shape is mostly enablement: training, councils, and a search for the people who lean in.
  • As champions appear, a growing share of the time moves to execution work with them.
  • By the end of the year, enablement still runs, but it is mostly fed by examples from the execution work.
  • Treat this as a typical shape, not a schedule. Your mix depends on where your people are starting.

In practice

At Rivian the two jobs fed each other. Training sessions were where I spotted the people who came back with questions and started building. Many of them became champions, and some became the builders on execution work. That execution work, redesigning workflows across capital programs and corporate real estate, is where the measurable savings and released capacity came from. Its results then became the examples we used in the next round of training.

Where it breaks down

The split is about measurement and staffing, not separate teams that never talk. Enablement without execution produces enthusiasm with nothing to show for it. Execution without enablement keeps every improvement waiting on a central team.

Try it with AI

Help me sort my organization's AI work into two categories:

- Enablement: getting the organization started (training, councils, champions programs) and finding people with the aptitude to pick up AI.
- Execution: delivering tangible return through workflow redesign, agents, and digital transformation.

Ask me about our current AI activities one at a time. Then tell me which category each belongs to, how each should be measured, where we're judging enablement on ROI or neglecting it, and who our enablement efforts have surfaced who should be pulled into execution.

Related