What 32 engineering disciplines taught me about change management

M. N. Dastur had been doing engineering since the 1950s. When I took on moving it to BIM-based virtual engineering, the scope was 32 disciplines — structural, piping, electrical, HVAC, and a long tail of specialisms each with its own conventions, its own drawing standards, and its own senior people who had been right about things for thirty years.

The software was the easy half. Here is what the other half actually consisted of.

Expertise is stored in the medium, not just the head

A senior engineer reading a 2D drawing is not decoding a picture. Decades of judgement are bound up in how that drawing is laid out — what gets shown, what is deliberately omitted, which conventions signal what. Ask that person to work in a model instead and you have not merely changed a tool. You have invalidated part of the apparatus their expertise runs on.

The resistance that follows gets labelled “change resistance”, which is dismissive and wrong. It is a rational response to being made a beginner at something you were the best at. Naming it accurately changed how I ran the programme: less training, more pairing, and a lot more listening to what the drawing had been carrying that the model was not yet.

Convert by discipline, not by project

The instinct is to pick a project and do it fully in BIM. It is the wrong unit. A project touches many disciplines at once, so a project-first rollout means every discipline is a beginner simultaneously, on a live deliverable, with a client deadline.

Discipline-first is slower to look impressive and far more likely to finish. One discipline converts, builds its own conventions, and produces people who can teach the next. By the time you reach the disciplines with the most entrenched practice, the argument is no longer theoretical — it is being made by their peers, in their own language.

The first converts should be respected, not enthusiastic

There is always someone eager to go first. Frequently they are eager because they are junior, curious and have little standing to lose. Their success proves nothing to the people you need, and can actively harm you: if BIM becomes what the young engineers do, it acquires exactly the wrong association.

Spend the effort recruiting a respected sceptic instead. Slower to win, worth several times more when won — because the question everyone else is silently asking is not “does this work” but “do people like me do this”.

Keep a bilingual period, and mean it

Every conversion programme is tempted to set a date after which the old way is forbidden. It feels decisive. In practice it converts a technical migration into a loyalty test, and drives the remaining 2D work underground where you cannot see it.

A deliberate bilingual period — where both are legitimate and the model is authoritative for defined things — costs more in the middle and finishes sooner. It also surfaces the genuine cases where the old way was still better, which exist, and which you want to know about rather than have concealed from you.

What actually convinced people

Not the demonstrations. What convinced people was clash detection catching a collision between two disciplines’ work before it reached site — the specific class of error that, in the 2D world, is found by a fabricator on a phone call, expensively and with someone blamed.

That was the argument, and it was not an argument about technology. It was: this stops the thing that makes your month terrible. Every conversion programme has an equivalent, and finding it is more valuable than any amount of training material. Ours took eight months to identify and about a week to spread once it had.