The technology is the easy part. The hard part — where most AI projects actually fail — is getting people to adopt it. Here’s how to lead the change in 2026. (dgm implements osFoundry as an independent partner.)

AI adoption is a people problem

It’s worth saying plainly: most AI projects fail on adoption, not technology. The tool works, but people don’t trust it, don’t use it, or were never trained. So change management isn’t a nice-to-have — it’s the difference between a working pilot and a dusty one.

Lead with the “why”

Start by explaining why — the problem AI is solving, for the business and for the people doing the work. People support change they understand and resist change imposed on them.

Involve the people who’ll use it

The teams who do the work know where the friction is. Involve them in choosing and shaping the use case. They’ll surface better use cases, spot pitfalls, and — crucially — feel ownership rather than imposition.

Address job fears honestly

AI raises real anxieties. Address them honestly:

  • Frame AI as augmenting people — removing repetitive drudgery so they focus on higher-value work — and mean it.
  • Be truthful about what’s changing; hollow reassurance backfires.
  • Keep humans in the loop on significant decisions (which also reflects UK ADM expectations).

Train for real use

Practical training matters: how to use the tools, what they’re good and bad at, when to trust output and when to check it, and the data rules (what not to paste into AI — see our AI use policy guide). The UK’s AI Skills Boost and other skills funding can help.

Show quick wins

Nothing builds adoption like a visible early win. A pilot that demonstrably saved a team time is worth more than any presentation. (See how to run an AI pilot.)

Where osFoundry and dgm fit

dgm treats adoption as change management, not just deployment. We build transparent, human-in-the-loop AI on osFoundry — where people can see what the AI did (audit logs), review significant outputs, and trust the tool because it’s not a black box. Because it’s source-available with a visual config editor, teams can understand and shape it, which supports ownership and training. Data control via self-hosting or an EU region (osFoundry publishes US/EU/JP regions, not a UK one).

dgm is an independent integration partner with zero integrations so far. To plan an AI rollout your people will actually adopt, book a consultation with dgm. General information, not specific advice.