The data maturity of companies: 3 major challenges remain to be met for the Business Lines to take full ownership of AI projects and generate value from them.
• 68% of respondents say that the average benefit from data and AI usage is increasing
• The average industrialization rate of AI use cases jumped from 27% to 46% in one year
• Only 29% of data departments manage their activity in correlation with business metrics
• 44% of respondents consider their data organization to be an obstacle
• Technical profiles are still highly sought after with 78% recruitment intention for Data Engineers and Machine Learning Engineers
Quantmetry, the leading French consulting firm in artificial intelligence, today unveils the results and lessons of the fourth edition of its qualitative study “Barometer of Data Directions 2022”, conducted in France with more than 50 managers and data managers from large and medium-sized French companies. The objective is to enable companies to analyze their data maturity compared to that of the market, to understand best practices in terms of organization, skills or technology, and finally to inform decision-making.
The data maturity of companies has become a major economic issue. Companies are looking for efficiency gains in their operations, want to improve their customer relationship or even develop new product and service offerings. In this context, spending on AI and data is increasing all over the world and this trend is confirmed in France: investments continue to grow, so much so that this year, nearly 8 out of 10 Data Directors consider that their company is investing up to the challenges of their data transformation.
In addition to investment momentum, organizations are also reaching maturity milestones. This year, the ability to translate their experiments into real AI in production, rather than remaining at the pilot stage, has greatly increased. The rate of industrialization indeed jumped from 27 to 46% in one year. Progress is largely due to improvements in strategy, organization and technological bases.
The next milestone for companies will be to ensure that these AIs are fully exploited within business processes. To do this, the companies surveyed in our Barometer placed 3 challenges at the center of their priorities.
Challenge n°1: Complete the organizational model
84% of companies adopt a hybrid organization with a central Data Office and data skills within the business entities. They are also the ones that are most successful in the industrialization of AI. However, the development of these models remains complex. Companies must align their organizational model to their architecture, which is why some are turning to Data Mesh, an organizational transformation aimed at decentralizing data management, while providing the consistency necessary for data to flow unhindered through the organization. Today, 44% of data departments still consider their data organization to be an obstacle.
“Historically, our Datalake operated in a hyper-centralized way. It was the Group that pushed the data to the subsidiaries. We have moved towards more decentralization to solve certain data accessibility issues. » Chief Data Officer
Challenge 2: Securing Skills Under Strain
38% of respondents believe that the skills they have are insufficient or poorly suited to their needs. Recruitment intentions therefore remain high, particularly for technical profiles, but companies are facing a shortage of talent which is a chronic problem year after year.
We see 78% of recruitment intentions on the profiles of Data Engineers and Machine Learning Engineers.
“The scarcity of Machine Learning Engineers makes them a priority target in our call for external resources. » Chief Data Officer
Challenge n°3: Bridging the cultural gap between data and business
The businesses do not yet fully integrate data into their management. Too often, data is not part of major strategic programs and businesses do not feel responsible for data ROI. Data departments are handicapped by overly operational management that is often unrelated to the value delivered to the business. Only 29% of them manage their activity in correlation with business metrics. For most companies, it is still difficult to comment on the quantified results of the industrialization of their AI and Data projects. The results are mainly measured on qualitative values and user feedback, but are truly encouraging.
“Continued support for businesses in terms of awareness and training is necessary so that they can become full owners of the models. » Maxime Havez, Chief Data Officer, Arkea
We wish to say thanks to the writer of this post for this incredible material
AI projects, three major challenges to take ownership of them | Viuz
We have our social media pages here and additional related pages here.https://www.ai-magazine.com/related-pages/