Why AI fails without company culture

AI adoption fails in 70% of cases due to cultural resistance, not technology problems. The three key factors: management sponsorship, hands-on training, and usage measurement. SMEs that treat AI as an operational priority see 3-5x higher adoption rates.

The graveyard of AI projects

In the Italian business landscape, there is a phenomenon that anyone working in the sector knows well but is rarely discussed publicly: the graveyard of AI projects.

These are systems implemented and never used. Internal chatbots that staff systematically work around. AI platforms adopted by management and ignored by operational teams. Pilot projects that never scale, year after year.

They do not fail because of technical problems. They fail because of cultural problems.

Why technology alone is not enough

An AI system, to generate value, must be used. And to be used, it must be integrated into the way people work daily — not just in formal processes, but in habits, informal conversations, and the ways people approach problems.

This requires a cultural shift that no technology implementation, however well designed, can impose from the top.

The research is consistent: 70% of digital transformation projects fail, and the primary cause is never the technology — it is resistance to organizational change. The data specific to AI is even more concerning: according to McKinsey, only 16% of companies that have implemented AI report achieving significant and sustained long-term impact.

The three cultural barriers

1. Fear of being judged by machines

In organizations where AI is introduced without clear and honest communication, people develop the perception that the technology is evaluating their performance, identifying their inefficiencies, and providing managers with arguments to reduce headcount.

This perception — founded or not — creates active resistance to adoption. People do the bare minimum with the system, find workarounds, and wait for the "trend to pass."

2. Threatened professional identity

In many Italian SMEs, the most experienced workers build a significant part of their professional identity on the tacit knowledge they have accumulated over time. They know how to manage the difficult supplier, how to handle the unhappy customer, how to fix the production issue that comes back every January.

When an AI system arrives that seems capable of doing some of these things, the natural reaction — even unconscious — is to protect that knowledge, not to share it to train the system.

3. Middle management as the blocking point

Middle management is often the most critical blocking point in AI transformations. Intermediate managers have structural incentives to preserve existing processes: their authority is often tied to knowledge of how things work, and an AI that automates part of their processes removes that power base.

The Doge Code: culture before technology

At IL DOGE DI VENEZIA, the first step of any AI transformation project is not choosing the technology. It is what we call "Phase 0": cultural buy-in.

Before installing a line of code or configuring a platform, we work with management and teams to answer three questions:

• Why are we doing this? Not "because AI is the future" — but why specifically this company, at this moment, wants to invest in this transformation. The answer must be concrete and shared.
• Who will be impacted and how? An honest map of impacts on staff, with a clear distinction between "it will change how you work" and "you will lose your job" — two very different things that are often confused.
• How will people be involved? AI is not installed on top of people. It is built with people. Those with process knowledge must be part of the implementation team, not passive users of the finished product.

How to build the right culture

Five operational principles that we have seen make the difference:

1. Start with volunteers: The first team to adopt an AI system should be composed of enthusiastic people, not people who have it imposed on them. Initial successes build the credibility that enables expansion.
2. Position AI as an amplifier: Internal communication must frame AI as a tool that amplifies people's capabilities, not one that replaces them. This is not spin — it is the reality in most SME use cases.
3. Measure and share results: When a team achieves tangible results through AI, those results must be visible to the entire organization. Internal success stories are the best change management tool.
4. Train, do not just inform: AI training must be practical and contextualized around people's daily work. Not generic courses on "what AI is" — but workshops on "how to use AI to do your specific job better."
5. Involve middle management: Middle management must become champions of the transformation, not victims. This requires redefining their incentives so that the success of the AI transformation aligns with their personal success.

The cost of cultural inaction

Every month of unmanaged cultural resistance is a month of missed ROI on a technology investment already made. In the projects we follow, the difference between an implementation with strong cultural buy-in and one without is typically 3-5x in results at 12 months.

Technology is necessary, but not sufficient. Culture is the multiplier.

If you are planning an AI project and want to understand how to address the cultural dimension in a structured way, talk to us. Change management is an integral part of the Doge Code.