AI adoption playbook

Weeks 1 to 4: foundation

Run a readiness check and share the results with leaders.
Hold a leadership alignment session on concerns, success measures, owners and the message about jobs.
Name the business owner and the approvers for each workflow.
Recruit champions in every team and agree their time with their managers.
Set up channels for updates, questions and error reports, using tools people already use.
Write honest answers to the questions people will ask, and brief every manager.
Send the first announcement.
Test the system against real business scenarios before anyone relies on it.
Agree the adoption and business measures, and record the starting baseline.

Before the pilot starts

Data access and ownership are agreed and documented.
Leaders have used the system themselves.
Champions are trained and have protected time.
The practice environment is ready, with realistic data that cannot affect real work.
Support channels are staffed and their hours are published.
Everyone knows how to report an error.

Weeks 5 to 8: pilot

Choose a pilot group that includes enthusiasts and at least one respected skeptic.
Train the pilot group in stages: demonstration, guided practice, then supported use on real work.
Check in daily in the first week, then weekly.
Track confidence and anxiety as well as usage.
Log every error and complaint as a test case.
In week 8, review the results with leaders and decide to roll out, fix first or stop.
Share what the pilot found and what changed because of it.

Weeks 9 to 16: waves

Start with the team that scored highest on readiness.
Give each wave intensive support for its first two weeks.
Have pilot users and champions help the next wave.
Share progress updates, including problems, every week or two.
Step support down gradually once each wave settles.

Measuring adoption

Measure the share of eligible work going through the new workflow, not logins.
Track how often reviewers change the system’s output, and spot-check approved items monthly.
Watch for the old process reappearing in spreadsheets or email.
Run a short monthly pulse on how confident people feel using the system.

If you are planning a rollout and want a second opinion on the plan, a discovery call with Kastling is a practical place to start. We build training and adoption into the project rather than adding them after go-live.