AI initiatives fail on adoption far more than on technology — so the discipline that protects your investment isn’t engineering, it’s change management. A playbook turns adoption from hope into a repeatable process. Here’s the playbook, and how dgm runs it. (dgm implements osFoundry, a separate company’s platform — we are not osFoundry.)

The playbook

  1. Involve users early. Bring in the people who’ll use the AI during scoping — people adopt what they helped shape.
  2. Communicate the why and what. Be clear about the goal, the change, and what it means for people’s work and roles.
  3. Design into existing workflows. Make AI make current work easier, not add a separate step.
  4. Train by role. End users, power users, and admins each need different training (see how to train your team on AI).
  5. Prove value with a quick win. A visible early result builds momentum better than any announcement.
  6. Address concerns transparently. Be honest about what AI does, doesn’t do, and where humans stay in control.
  7. Measure adoption. Track actual usage — not just that it was deployed.

Run it throughout, not at the end

Change management isn’t a final phase — it’s woven through the project. Involve users in scoping, design for their workflows in the build, prove value in the pilot, and measure adoption after launch. Each phase is an adoption checkpoint (see change management for AI adoption).

Measure adoption, not deployment

The key metric isn’t “did we launch it” but “are people using it.” Track usage by intended users, the share of relevant work going through the AI, and satisfaction. A system that’s live but unused isn’t delivered — and adoption metrics tell you whether the playbook is working.

How dgm helps

dgm builds this playbook into every engagement — involving users, designing for their workflows, training by role, proving value in phases, and measuring adoption — as part of the $3,999/month implementation (after a $399 assessment), not as a separate line item. If you’d rather explore the platform yourself first, go straight to osFoundry; if you want AI your team actually adopts, that’s where dgm comes in.