Companies attacking the AI industrialization cycle are not without management problems. Data Mesh-backed federated governance offers perspectives that need to be validated, said Judge Olivier Malle, director of the Data practice at Cognizant.
>> This article is taken from the Alliancy Insights download guide “Data and AI: Prioritizing Impact and Scaling”.
What big trends are you seeing in terms of AI adoption in business?
Olivier Malle: After the first wave of AI in 2013 and 2014, which was characterized by numerous PoCs, organizations are entering the second phase of modern AI. This is the cycle of industrialization. This is a real challenge today.
However, this requires the solution of rather deep problems, for example, those related to management. And this is a project that takes time to set up and complete. Moreover, in my opinion, maturity in France is still quite low in terms of AI at scale.
See also: Marie Gepel, Technip Energies: “The future is in the development of data ecosystems”
Considering that one of the prerequisites is operational management, I would say that this applies to only 10% of French companies. Even fewer have a complete structure: management, quality and architecture.
We are still at the beginning of the adventure. The stakes are huge in terms of the associated cost. But companies have realized that this value creation requires preconditions.
Data Mesh seems to rehabilitate control. Do you see a better consideration of this element of the data strategy?
In fact, it varies greatly from company to company. Certain sectors historically heavily regulated, such as banking and pharmaceuticals, verticals that we at Cognizant know very well, are leading the way in this area. In particular, they have created cross-functional structures to manage quality and compliance issues. This does not mean that they are at the end of the road.
On the other hand, many actors still do not understand the problem. Large companies do not. They all appointed a CDO who deployed the management policy and generally defined the roadmap and vision. However, some blockages may appear in the operating deviation.
The idea behind federated governance is to bring flexibility to these processes while better respecting the natural organization of companies. It is hoped that Data Mesh will solve various problems. We have also implemented it with some of our main clients, in particular with the Americans. In France, the concept is still mostly a buzzword. Projects have been launched, but we must wait for the results to judge its implementation.
Federal governance is accompanied by a cultural transformation. Should we operate in the Big Bang mode?
This culture, it can be changed gradually. However, some fundamentals of this culture cannot be changed. In a company built around centralization and with employee practices structured around this approach, it would be illusory to think that it could be destroyed. You must come to terms with this. This will undoubtedly require partially centralized governance structures.
In a younger, more agile company, it’s easier to take advantage of a clean slate to invent new management.
Could the involvement or lack of business lines still be a barrier to data and AI adoption today?
This participation is directly related to maturity and culture. The regulated sectors have integrated this responsibility into processes and organizations. On the other hand, for companies where data has always been considered a by-product of IT, awareness is more of a challenge. Removing this barrier requires pedagogical and communication work on the Data value contribution.
As for the second aspect of your question: are professions ready to move to AI? A company’s recruiting strategies are the lever of action. The challenge here is to infuse the technological capabilities of AI with the business environment.
More generally, the current obstacles are no longer of a technical nature. There are many tools, maybe even too many. The strongest brake is the maturity and alignment of IT, business and management levels. The latter also plays a role in understanding and allocating the right budgets.