Loading...
Loading...
Digitalisation, AI and organisational change
The ERP went live months ago. Much of the work still runs on spreadsheets.
It happens often, and almost never for a technical reason.
Transformative intelligence is an organisation's capacity to evolve as a whole and in a coordinated way: its people and its structure, its processes and flows, its systems, its skills and its training.
It is not change handed down from the top. It is the genuine involvement of people at every level, and their appetite to push: the ability to adapt for real and to drive change, rather than simply absorbing it or, worse still, having it done to them.
Organisations change because everything around them changes: markets, the technologies that come within reach, the rules, the skills required. Choosing a system and installing it is the part you can plan and control. The way people work, by contrast, does not change because a project plan says so.
This is, in all likelihood, the organisational capability that leadership should invest in ahead of any other. And it is not confined to one sector: the pace of change is high in every market, whatever the business model.
Not one area at a time, but all of them together and in the same direction. That is what makes this hard to do, and what makes it decisive.
The name can mislead, and it is worth clearing that up straight away. Transformative intelligence is not a technology and not a category of artificial intelligence. It is the organisational capacity that allows AI — or any other system — to have a real effect on everyday work. Technology is bought; this capacity is built.
A project delivers a configured system and documented processes, and it trains the people who will have to use them. These are verifiable results and, in most cases, they are achieved. What a project cannot deliver is the evolution of the organisation receiving it: roles that are redrawn, responsibilities that shift, skills that mature, people who start putting things forward instead of waiting for instructions.
When that part does not happen, the system goes live and the company carries on working as before, with a new tool.
This is the heart of the matter, and “involvement” is the most abused word in it. Involving people does not mean informing them: however well an announcement is written, it lands once the decisions are already taken, and the people receiving it can tell the difference. Real involvement shows in four things.
People are brought in while the decisions are still open.
The people who will use the system take part while there is still something to decide. After that it is communication: useful, but not the same thing.
Objections change something, at least sometimes.
If observations have never once changed a decision, people learn that fast and stop making them. Participation, at that point, becomes a formality.
Proposals come from any level, and they go somewhere.
The people closest to a process see what is not working before anyone else. Either improvement is something everyone does, or it does not happen at all.
Scepticism is treated as information.
People who raise objections have often seen something before anyone else. When a sceptic stops speaking, you have not solved a problem; you have lost a source of information.
When these four things happen regularly, adaptation stops being an initiative and becomes a habit. That is the point at which an organisation begins to change on its own.
Six recurring situations. When three of them appear together, the issue is usually not the tool that was chosen.
The system is live, but much of the work still runs on spreadsheets.
Decisions are taken a long way from the people who will do the work every day.
Doubts do not surface in meetings. They surface afterwards, through informal channels.
New technology arrives, but roles and responsibilities stay exactly as they were.
Improvement proposals only ever come from the top.
Six months on, it is hard to point to what has actually changed.
When it reaches only operational levels, change halts precisely where decisions begin. Those who lead need to understand the technology at least as well as those who use it.
An org chart that stays identical while tools and processes change is the most reliable sign that change stopped at the door.
The exception, the edge case, the reason one override is approved and another is not: none of it disappears because software arrives. Everything else gets a named owner, or it comes back later as a parallel spreadsheet.
The useful question is about everyday work: do people work differently from before?
Bringing artificial intelligence into processes makes the question more pressing, for a specific reason: unlike an ERP, an AI-based system changes the content of the work and not just the tools you do it with. It shifts the line between what a person decides and what a model proposes, and with it the skills required.
An organisation that does not evolve its roles, its training and its leadership adopts AI as an accessory: it ends up in the hands of the people who already knew how, and nothing moves where it was needed most.
You will find it everywhere, and it is the first number anyone quotes on this subject. It comes from an estimate its own authors called “unscientific” (Hammer and Champy, 1993). A peer-reviewed study later traced its most-cited sources and found no valid and reliable empirical evidence behind it (Hughes, Journal of Change Management, 2011). We note it because we will not be using it.
Two meta-analyses have examined this: He and King (Journal of Management Information Systems, 2008) across 82 studies, and Schermann and Merz (ECIS proceedings, 2018) across 226 studies and 42,330 IT projects. Both find that involving the people who will use a system improves outcomes, though with a weak-to-moderate rather than decisive effect. The title of the second says the essential part: participation is necessary, but not sufficient. As far back as 1948, a field experiment (Coch and French, Human Relations) found productivity after a change to be directly proportional to the degree of participation, with turnover and conflict moving the other way.
Some 76% of Italian SMEs have not invested in artificial intelligence and do not plan to, and only 7% have started structured training programmes for their staff (Osservatorio Innovazione Digitale nelle PMI, Politecnico di Milano, May 2026). The two figures need reading together: most Italian SMEs have not yet touched AI, and the share that has touched skills is smaller still. When the investment does come, it will land on organisations that have not prepared to receive it.
The concept used to study organisations capable of reconfiguring themselves is dynamic capabilities, set out by Teece, Pisano and Shuen in 1997 and taught on every MBA. It describes precisely how a firm recombines resources and competences when its environment changes, but it says very little about what happens to the people while it does. That, in our view, is where the difference is decided.
The boundary
If change leaves people behind, it is not a transformation: it is a restructuring.
These are two different movements, and they deserve two different names. An organisation can reconfigure itself very effectively while losing half of those who made it work: it happens, and sometimes there is no alternative. But it is another operation, with other costs and other results — and calling it a transformation does not make it one.
Gitogi does process digitalisation: Odoo, AI integration, custom software. We are accountable for that work — for the schedule, for the quality, for the system doing what it was designed to do. What no supplier can put in a contract is that the organisation will work differently afterwards. That part belongs to the people inside it, at every level.
We talk about it because we hold our work to a single measure: in the end, the client should no longer need us. An organisation able to evolve on its own no longer has that need.
Autonomy.
The idea lives in concrete projects. These are the three places where, in our work, you can see whether an organisation is really evolving.
If you are weighing up a new ERP, or bringing AI into your processes, the useful question comes before the tool: who will use it, and how they get there.