How can large organisations understand what digital skills their employees actually have without relying on time-consuming manual assessments? A research article in Information Systems Journal analyses an organisation’s work to introduce an AI-enabled platform that automatically infers employee skills, supporting the company’s digital transformation.
The study follows how the company Johnson & Johnson developed and deployed a so-called Digital Talent Platform, which is a system that uses artificial intelligence to analyse existing HR data and other digital traces in order to infer employees’ proficiency levels across a set of defined skills. The aim is not to monitor performance, but to provide a basis for talent development and strategic workforce planning in a global organisation with more than 130,000 employees.
”Many organisations struggle to envision and build a digital-first workforce because they lack a clear understanding of their employees’ real skills and the future skills needs”, says Olgerta Tona, Associate Professor of Information Systems at the University of Gothenburg and co-writer of the article.
The researchers show that the technical solution is only one part of the effort. Equally important is how the organisation prepares for and manages the questions that arise when AI is applied to employee data.
”What is important to understand is that deploying AI requires more than just having the right tools. It is equally important to have strong organisational routines and practices that actively engage employees in the process”, Olgerta Tona points out.
The study identifies three key organisational practices that were critical for making the platform work in practice. The first one is blueprinting the future workforce. The company began by developing a shared view of which digital capabilities and skills would be needed going forward, based on business strategies and with input from subject matter experts.
The second one is managing ethical data work across borders. The deployment required extensive work on data protection, local regulations, engagement with workers’ representatives, and ensuring that employees could opt out.
The third crucial practice is compensating for AI blind spots. Employees and managers were given opportunities to review and adjust the AI-generated inferences, update their own data, and understand how the results would be used, which increased transparency and trust in the system.
According to the study, the introduction of the platform is associated with increased internal mobility, better matching of employees to digital roles, and lower voluntary attrition in digital and data-related positions. At the same time, the authors emphasise that AI solutions of this kind are not “plug and play”. They need to be rolled out iteratively, with careful attention to ethics, legal requirements and employee experiences.
The study thus offers concrete lessons for other large organisations that want to use AI to understand and develop the digital skills of their workforce without compromising integrity, transparency or trust.