Updated 30 October 2026

Training explains a tool. Whether people are still using it six months later depends on something else: whether the new way of working holds up in their real work, who helps when something goes wrong, and which old steps actually disappear.

Start with the process, not the tool

Licences and access can be handed out in a week. Then the usual pattern follows: a few enthusiasts get better, most people try it once and go back to what they know. That isn’t resistance to change. People have a full day of work and the new tool hasn’t taken anything off it yet.

So we first ask which work should be different. Pick one or two processes that matter to leadership and that repeat. Name who does them, what the input is and how you’ll know they have improved.

Work on people’s own tasks

Our workshops use the flipped classroom method. Instead of a tour of features, people bring a complaint, a quote, a spreadsheet or a document they are really dealing with that week. They try the solution themselves and leave with a method they can use the next day.

We form groups by type of work, not by position in the organisation. A salesperson and an accountant have different routines. A joint session is too general for both.

What the data from one manufacturer showed

At Isotra, a Czech manufacturer of shading systems, we worked this way with around 120 people. Five to eight months after the workshops we surveyed the first group on how often they used AI. According to their responses, the share using it several times a week or daily rose from 17% to 78%.

The same data also shows the limit. AI helps most with emails and copy. Only 39% agree that it lets them do their work to a higher standard, and four out of 23 saved no time. Workshops change individual habits. Changing operations takes further steps. We have published the full study, including its method and limits.

Test under normal conditions

A process that works in a meeting room may not work in a warehouse or out in the field.

Illustrative example: purchase approval works on a computer, but the manager needs to decide from a phone. Without a mobile route, the work goes back to email. Banning the old channel won’t fix that. First you need a usable new process, a way to handle exceptions and a way to record emergencies.

Agree ownership and the switch-over

Before go-live, answer four questions:

  • Who is accountable for the result of the process after the change?
  • What will people stop doing? If nothing, you have added to their work.
  • How are exceptions handled and who do people turn to?
  • When can the old process be safely retired and what happens during an outage?

Measure work, not attendance

The number of people trained says nothing about change. Measure use before and after, ideally from tool logs and backed up by a survey. Alongside that, track cases handled, workarounds, quality and time to resolve. Agree the baseline in advance, because a retrospective estimate is weaker evidence.

You can’t draw conclusions about a whole company from an enthusiastic team. A pilot needs people with different experience and working conditions.

How we approach this at ENTER

Adoption is the second phase of our work and follows the Audit, where we choose processes with leadership and agree the baseline. In Adoption we work with teams on their own tasks, name the owners and measure with the same yardsticks after the change. What people keep doing by hand becomes the brief for Integration, where the change moves into connected systems.

Related: Isotra case study · Workshops · How we work