Four things that get called “AI enablement”
Buyers use the same phrase for four different purchases. Being clear about which one you need saves a wasted quarter.
- Training teaches people what a tool does. It is useful when the work is already well understood and the gap is familiarity. It rarely changes behaviour on its own, because nothing about the work itself has changed.
- Tool rollout gets licences assigned, settings configured and announcements sent. It is necessary, but it is an IT project. Usage dashboards go up for a few weeks and then flatten.
- Leadership workshops help executives decide what to invest in, what risk to accept and how to govern it. That is a decision-making engagement for a small senior audience, and I run it separately through dipankar.org.
- Enablement, as I use the word, changes how a specific team does a specific piece of work, and leaves someone inside your organisation able to keep it running.
This hub is about the last one. It borrows from the other three where useful, but the unit of work is always a real workflow with a real owner.
What an enablement pilot actually involves
A pilot starts small on purpose: one team, one or two workflows that recur every week, and an approved tool.
Baseline before change. I sit with the people doing the work and map it as it runs today: where the inputs come from, what is read, drafted, checked and sent, and where rework happens. We record how long it takes and what “good” looks like before anything changes. Without a baseline, a pilot can only report feelings.
Data boundaries. For each workflow we write down which information may go into which approved tool, which may not, and where output must stay inside a system of record. This is usually one page, and it is the thing staff most often say was missing from earlier rollouts.
Review points. Somebody accountable signs off AI-assisted output before it leaves the team: a finance reviewer, a support lead, a client partner. We agree who, at what step, and what they check. Human review is part of the design, not an apology for it.
Coaching on real work. Short sessions on the team’s own tasks, then working alongside people as they use the new workflow, then a weekly check on what is sticking and what is not.
Measurement. Repeat use by the people who should be using it, time per item, quality as judged by the reviewer, and rework. Licence log-ins are not the measure. The adoption measurement resource shows the approach.
The signature deliverable
At the end you hold an AI enablement diagnostic and pilot: the baseline, the redesigned workflow, the data boundary and review-point record, the coaching material the team actually used, the measurement results, and a go, revise or stop recommendation for wider rollout. Anonymised pilot output is included so a sponsor can see what changed rather than read a summary of it.
How acceptance is judged
The sponsor and the internal owner agree the measures before the pilot starts. Acceptance means the measures were collected and reported honestly, the owner can run the workflow and coach a new colleague without me, and the recommendation follows from the evidence. A pilot that ends in “stop” is an accepted pilot if the evidence supports it.
Ownership after the pilot
The workflow, its playbook and its measure belong to a named person in your team from the first week. I coach them through the pilot so the handover is a formality rather than a cliff edge. If nobody can be named, that is a finding in itself, and the fractional enablement lead route may be the better start.
Choosing the specific route
The comparison above maps each route to the situation it fits. A few common confusions:
- Unused licences after a rollout point to adoption rescue. Many ownerless pilots point to programme rescue. They need different first assessments and different sponsors.
- A specific department (finance, support, operations and so on) is served by business-team enablement, which links to the department pages. Software teams are served by engineering-team enablement.
- A platform decision or rollout (ChatGPT, Microsoft 365 Copilot, Google Workspace) is covered under AI tool rollout.
- If you are still deciding whether you need a course at all, start with AI training versus enablement or the free readiness assessment.