Helen Poitevin, Gartner Research Vice President, interviews Gartner about the status quo of AI in human resources, how workforce’s concerns can be addressed, and how artificial intelligence can contribute to a better work environment.
Around ten to 15 percent of all HR departments worldwide already use artificial intelligence, estimates Helen Poitevin, Research Vice President at Gartner. The American IT research and consulting firm has identified three use cases of AI in human resources: Talent acquisition, Voice of Employee and virtual HR assistants.
How is artificial intelligence currently used in HR departments?
On the one hand through Conversational Agents, who are directly assigned to the applicants. They help to find a suitable employment opportunity for the candidate. In addition, they contribute to the evaluation process with appropriate questions as to whether the applicant is suitable for the company.
We assume that such Conversational Agents improve the applicant’s experience with the company: Many questions can be answered in an automated process and many candidates even prefer this type of conversation because they don’t have to deal with how they affect their interviewer.
The second use case is about optimizing recruiting processes in several respects: on the one hand, it is about increasing the availability of information about the candidate for the executive recruiter, and on the other hand it is about automating appointment arrangements, which is particularly important for companies in which 200 positions have to be filled in one day.
The third use case is fundamental and probably the most difficult to achieve. It is about a standard anthology and an understanding of jobs and work experience and how they relate to each other. The goal is a better understanding or machine matching of candidates and job descriptions. Some providers are already working on developing a kind of skills anthology or taxonomy and automating its recognition. In this context, labels and comments on this information in job descriptions and applicant information are also important. However, this is still a major technical challenge.
Why isn’t every human resources department already working with artificial intelligence?
Worldwide, I estimate the number of companies that use artificial intelligence within HR and Recruiting at about ten to 15 percent. Rather 15, because many are not even aware that their recruiting software already uses artificial intelligence. Especially the large providers of such systems have already integrated elements of AI. The use of artificial intelligence in human resources is expected to spread rapidly, partly because it optimizes and automates user experience and processes.
What objections do those who do not yet work with it have?
Although to my knowledge there is no application where artificial intelligence makes the final decisions autonomously, there are certainly such fears in the market and from government agencies. Especially in Western Europe and North America, there are basically more concerns about technology bias. Such fears are certainly not completely unfounded, but no one is currently trying to develop such functions. Many current use cases are very far away from fully automated decisions.
At what point of development are we right now?
Artificial intelligence is omnipresent in our office systems, think of applications, spreadsheets or documents with suggestions for visualizing data. Many HR teams work with LinkedIn or Jobsights and on the US market with the Google Job API, which is equipped with AI capabilities.
However, we tend to see artificial intelligence as a super-intelligent machine that works autonomously. But it is much broader: AI is a set of skills that can be used in a variety of ways. Many systems already work with quasi artificial intelligence in the form of predictive insights, which is certainly the most common form of adaptation. We also see a growing use and interest in chatbots used by candidates.
What challenges do HR professionals face when it comes to using these tool sets, especially Voice-of-Employee programs?
Artificial intelligence in VOE programs can use those models and analytics that are accurate enough to locate the correct schemata in long text responses. The adaptation of these tools is a kind of work in progress and there still seems to be the idea that there is a magic button somewhere that you just have to press to get a mood barometer about what is being said in the company.
You need to understand which data sources and tools help you most to capture moods, ask the right questions, and analyze more open answers than the same models in Engagement Surveys.
What about the employees’ scepticism about surveillance?
Part of the challenge is certainly adequate communication. The fears of the employees vary from market to market, in Western Europe and North America the fears are certainly greater than in China, India or Singapore. By the way, some of these tools enable managers who are currently undergoing change processes to gain insights into their employees’ fears, not to track or pass them on, but to have more open discussions with those affected by the changes. It’s about listening and communicating back.
What threats could VOE programs pose?
The greatest danger of VOE techniques is over-listening and not being able to respond appropriately. You have to keep in mind how much you ask and what response would be appropriate. Many executives and organizations are unfortunately poorly equipped for this because they cultivate a culture of secrecy or are in another situation where open communication is not possible.
However, fears about how the data is collected, who has access to it and that employees will be punished for honest feedback have always existed in employee satisfaction surveys. These fears don’t change with AI, it’s more about how we get organizations to listen more and better. This requires communication and above all transparency.
What significance will virtual assistants have in the future?
We expect virtual assistants to further develop self-service information and automate the answering of HR-related questions. Of course you have to train these bots and make sure that everything the bot is not able to do is forwarded to a human.
In general, however, we see that many HR executives are interested in virtual assistants also from the perspective of employee experience: Most employees in a company have very little contact with the HR department and it can be difficult to find out where to go when looking for information on a specific topic.
What skills are expected of the personnel of the future when they work with bots, agents and assistants?
The most important skill for HR professionals is adaptability. People really take a long time to adapt to, with and thanks to technologies that they use around them. HR also urgently needs to work on their digital skills and agility to gain an insight into how technology can help them.
Many HR professionals like to focus on their ability to interact with others. What matters now is adaptability and a certain learning agility to discover new things, experiment and find new ways with the new tools that are given to them.
Can artificial intelligence also contribute to improving the working atmosphere and the well-being of employees?
There are some tools. One of them and this sounds frightening is the identification of burnout or psycho-social risks. Let’s assume that AI could investigate in social media streams if someone is depressed or has suicidal thoughts, and then connect that person with experts who can help him through this situation, then of course that would also benefit the work environment.
Such functions are currently still at a research stage and will certainly first be used in high-stress environments and hazardous work situations where employee safety and health can become a critical factor.
Another lever is the approach of automating boring work. We often talk about the “bad” parts of our work that we would like to get rid of and the “good” parts for which we get recognition and the “great” parts that are fulfilling and of which we would like to do much more.
We already use AI to automate the bad shares, which enables us to have more time for the good and great work. The interactions of bots and artificial intelligence in general are seen by more and more employees as time savers and stress minimizers. It can’t always be that way but sometimes it works.