For an optimization of the potential of AI in business
Data is a mandatory starting point when looking to build an AI solution. The work of humans, AI is nothing more than what we decide to put into it. Here is an analogy that will simplify the matter: today we are able to follow the recommendations of a GPS because hundreds of routes are listed there. Listed by men, for men. However, despite the intelligence of the tool, the driver’s experience allows him, in certain situations, to disregard the recommendations of his GPS. To transpose this image to the corporate world, we could state that AI must adapt to each context, and therefore each company. But AI in HR is tricky and can’t be too scientific. It is experiences and interpretable data that feed it. Dependence on the availability of interpretable data has, until now, hampered the development of AI solutions.
The challenge for a company is to have control over the goals and trade-offs of AI solutions, unlike Youtube for example, which is a huge Black Box. Within ten years, it should be possible to process qualitative and textual data (skills, jobs, CVs, profiles, careers, departments, learning and development, etc.) and their relationships (skills on a CV and level of mastery, skills by role , links between learning and career paths, careers in certain departments). The first step to optimizing an AI solution for HR is to transform qualitative career and employee data into interpretable, measurable and comparable data points. The second step is the formulation of complex models that allow meaningful predictions to be made about specific actions or career situations, based on this data.
It’s not up to organizations to adapt to AI
AI in HR is not meant to make decisions for HR teams because it is too complex. The interpretation of HR data imposes additional problems
- Complexity: in order to extract information about the experience of employees, their skills and their career, the context is of paramount importance. Take, for example, the case of an employee whose job title includes the characters “DR”. Does this mean that the employee has a doctorate? A doctorate in medicine? Is that the abbreviation for director? Is it an acronym?
- Consistency: similarly, the consistency of certain terms is essential. For example, if someone says they have “sales” skills, that doesn’t always mean the same thing. Is it B2B or B2C? Does its use change over time (eg junior vs senior sales position) and space (eg different company/different types of services)?
In companies of all sizes, HR use cases cover all aspects of their work: streamlining recruitment processes (automatic screening, appointment scheduling, onboarding), talent management (career guidance, professional training ), strategic decision-making (skills mapping, search for experts, reorganization). However, HR departments are ill-prepared to gather and process the data needed to build an effective AI-based system and therefore have to hire AI services from third-party vendors. Collaboration with humans is a key issue for enterprise solutions, the progress of which will positively affect the reality of work.
The technological maturity of AI will humanize work
Opaque in its operation, AI is an organic tool that requires an interaction similar in some respects to that between humans. In 10/15 years, we must be able to offer customizable AIs, which adapt to the goals and objectives of each organization. The diversity of companies and use cases makes it difficult to integrate AI systems into a company’s functional architecture, which is usually based on existing business processes. Done right, the application of AI for HR goes beyond process-oriented and transactional approaches and enables HR to be truly transformative to accelerate the growth of people and their business. Despite its complexity of realization, AI for HR has all the potential necessary to really transform the company: it is mainly for this reason that AI must adapt to the thinking of the user. In the long term, this will undoubtedly lead to steering the company through skills: having a vision of a pool of skills at the moment T, and confronting these trends with a growth challenge.