The risks and rewards of AI deployment
Currently, most HR teams use AI in high-impact, low-risk applications, such as CV screening, HR ticketing, personalized learning, and chatbots that respond directly to employees’ questions about pay and benefits. Most of this is supplied by major software vendors such as Workday or Oracle, as part of their HR offerings.
With respect to learning, for example, AI can create personalized plans and indicate suitable career paths based on an individual’s strengths and weaknesses, and the projected future needs of the business.
“If HR leaders can get comfortable with the risks, the next most fruitful AI use case would be performance reviews,” suggests Garr. “Some vendors already offer this, and HR leaders are beginning to get excited about it.”
To assist with performance reviews, the technology can pull information from multiple sources, including manager notes, peer comments, and quantitative performance measures such as sales volume, customer feedback like net promoter score (NPS), or task completion. It can then summarize this information and produce a first draft of the review, which a human manager could check, amend as required, and then approve, with considerable time saved.
However, biased, incomplete, or inaccurate information could lead to an unfair review. This could, in turn, demotivate those affected or, in extreme cases, even result in a legal dispute. Strong governance of the technology is, therefore, imperative.