Artificial intelligence (AI) and data science were hot business subjects in 2024 and will remain in a burner before in 2025, according to Recent research published In IA in actiona series of columns focused on technology in the Mit Sloan Management Review.
In Five AI and data science trends for 2025Tom Davenport and Randy Bean researchers describe the ways in which AI and our data -based culture will continue to shape the business landscape during the coming year. Information comes from a range of recent IA -based research projects, including 2025 The reference survey on AI and data leadershipAn annual survey on data, analysis and managers of AI led by the Bean, Data & IA Leadership Exchange educational company.
The five trends range from the promise of agentic AI to the struggle on which the C-Suite role should supervise the data and responsibilities of the AI. At a glance, they reveal that:
The leaders will face both with the promise and the beateering around agentic AI. Agentic AI – which manages tasks independently – is increasing, in the form of generative AI robots which can perform certain content creation tasks. But the authors say that it will take a while before these tools can manage major tasks, such as making a travel reservation or carrying out a banking transaction.
The time has come to measure the results of the generative experiences of AI. The authors say that very few companies carefully measure the productivity gains of AI projects, in particular when it comes to determining what their knowledge -based workers make time released that these projects provide. This is essential to take advantage of AI investments.
Reality on data from data is installed. The authors found that 92% of respondents in the survey estimate that the challenges of cultural management and change are the main obstacles to data creation and AI – indicating that the transition to AI is much more than technology.
Unstructured data is again important. The possibility of applying generative AI tools to manage unstructured data, such as text, images and video – puts a renewed accent on the form of all this data, which takes a lot of human efforts. As the authors explain that “organizations must choose the best examples of each type of document, tag or graph the content and have it loaded in the system.” And many companies are simply not there yet.
Who should perform data and an AI? Expect a continuous struggle. Should these roles be focused on the commercial or technological side of the organization? Opinions differ and that the roles themselves continue to evolve, the authors say that companies should expect to continue to fight with responsibilities and report structures.