The subjects of data valorization and exploitation of artificial intelligence are discussed within the Executive Committees of the largest companies. In France, the public authorities also intend to play a leading role. They are launching phase two of the national strategy with the ambition of spreading AI across all organizations.
Maturity has progressed, but much remains to be done, particularly in the area of human capital. Technological mastery is not enough to transform a company into a Data Driven Company.
EDF map of professions and expected skills
On the HR and training aspects, different approaches are implemented, as experts testified during the last Data Circles meeting organized by DataScientest. So, even before recruiting or training, it is still necessary to identify the right skills to bring together, both within the Data sector and the business lines.
EDF has therefore undertaken the creation of a nomenclature of data professions, including the Data Analyst, the Data Scientist, the Data Architect, etc. These profiles are associated with expected skills.
Didier Canon, director of the EDF Digital Academy, was however able to note that depending on the profession, expectations could fluctuate. The energy company has therefore strived in recent years to structure Data professions and their definition. The list now includes 11 professions.
The company is now working to establish the exact missions for each person, and to associate skills (technical and transversal), training and career paths, as well as levels of expertise.
This position-based approach can, however, run into difficulties, as with the Data Analyst function. Often in the profession, this Data Analyst position requires specific skills. And these can be brought together by an employee without their job title being that of Data Analyst.
For Didier Canon, it is therefore appropriate to prioritize the development of skills among the workforce. What about internal professional retraining? “It’s a relative failure” at EDF, which had opted for work-study training of 12 to 18 months. Too long for group entities seeking to fill internal positions.
Reskilling for Data skills in the professions
To meet its skills needs in the Data professions listed, EDF has designed associated training courses. The critical point in this approach then consists of selecting the right input profiles. And prerequisites are essential. This starting capital makes it possible to estimate the effort required to strengthen or acquire the target skills.
Valérie Goutard, director of Societe Generale’s reskilling program, identifies similarities with EDF: development of professions, multiplication of employment codes, listing of skills, anticipation of new professions, etc. Maturity is progressing on this understanding of profiles and skills Data, by nature transverse.
“Data is everywhere. She is in all professions. She must even be in contact with the profession to understand its needs,” underlines the bank spokesperson. And this understanding therefore requires skills, acquired in particular through training and recruitment.
“We recruit Data profiles, just out of school, but also expert level candidates. We also try to save money. Recruitment being more tense, we are developing alternative solutions.”
Among these alternatives is reskilling (a term preferred to retraining, which is less well received), developed through a program entitled “Reskilling Initiative”, and involving 200 to 250 people per year. And Valérie Goutard is pleased with the results recorded, even if the approach works differently depending on the profession.
Choose your battles for greater efficiency
Data Quality Manager and Data Designer pose little or no difficulty. The profiles entering training, mainly from IT, bring together 60% of the expected skills. The remaining 40% is acquired during a 6-month training cycle.
It is more complicated, however, for the Data Scientist position. So, no promotions here, but individual career paths launched only when a position is opened by a manager. As for the Data Analyst, as at EDF, he is the subject of particular attention. This involves defining its contours more precisely.
For Valérie Goutard, reskilling has certain advantages, but on condition that you choose your battles, that is to say, concentrate on a few professions. The HR expert considers reskilling to be a “very good complement” to recruitment. A source of internal social consensus, reskilling is appreciated by employees because it is beneficial in terms of mobility and employability.
Please note that candidates for the Société Générale program are signatories to training waiver clauses. This is also the case at Ceva Santé Animale. For its Chief Data Science Officer, Thomas Lewiner, training, including with a reskilling perspective, has various advantages.
In particular, it constitutes a means of democratizing and disseminating Data skills in the professions. For this, Ceva Santé Animale trains in particular employees “in contact with the productive part and customers” in Data analysis and Data Science.
Livestock technicians, salespeople, veterinarians, and even customers, were trained. The selected candidates have in common that they are “very motivated”. And this approach leads to good results, appreciates the Chief Data Science Officer.
These upskilled profiles are useful for contributing to Data projects carried out in the organization. They also contribute to improving overall maturity. Other skills are also necessary. Ceva Santé Animale also offers less in-depth training with the aim of helping employees in their empowerment and use of data.
A Data school for internal and external students
Thomas Lewiner notes that these training courses allow “to prepare the ground and bring out profiles”, some of whom will access DataScientest training courses to develop in-depth skills.
If Societe Generale is experiencing difficulties in certain professions, including that of Data Scientist, Thomas Lewiner observes them in positions related to data governance. These professions are of greater interest to employees with an executive profile, recognized authority and generally close to IT.
In order to advance skills and maturity, the La Poste group decided in 2022 to create its own Data & AI school, open to work-study students and employees undergoing retraining or reskilling.
Four priority functions have been identified by the company: Product Owner Data, Data Analyst, Data Scientist and Data Engineer. In the Data Analyst profession, also central to La Poste, a fifth promotion is launched. Each has 8 to 15 participants.
To attract and industrialize, La Poste relies on certification training and the incentive to recruit internally. Candidates also follow work-study courses. Finally, the group gives a central place in its courses to responsible AI and ethics.
“Our goal in 2024 is: one promotion for each function – Product Owner Data, Data Scientist and Data Engineer – and potentially two from Data Analyst. 2024 will be a reference year for launches,” declares Tatiana Meunier Audap.