UKG integrates generative AI into HR tech with Google Vertex
UKG is using Google's Vertex AI platform to test and explore generative AI and sees it delivering conversational capabilities across its services and internal processes.
HR vendors have yet to roll out generative AI in a significant way. The early applications are for tasks like job description writing. But UKG, a large HR tech vendor, believes generative AI will change many processes and make them conversational.
Take customer support tickets. UKG used generative AI to analyze a year's worth of tickets to identify patterns. A pointer to the correct documentation may fix a problem, but others may take an engineering effort.
Then there are recurring problems noted by only ten customers over a year -- "a pattern that was harder to discern," said Hugo Sarrazin, chief product and technology officer at UKG. But revealing the pattern enables an automation fix, he said.
"We're going to be able to make more things self-service," Sarrazin said. He said there would be fewer tickets because there would be a documented path to fix the problem. The customer would get help "at the moment of pain."
UKG is still testing and exploring generative AI using Google's Vertex AI platform. The firm said the big changes in user experience, including more conversational searches and access to information, will arrive in November. According to Sarrazin, in an interview, it will have a significant impact.
UKG is a private firm formed in 2020 after the merger of Ultimate Software and Kronos Inc. The merger created one of the larger HR tech platform providers, reporting $3.65 billion in revenue in 2022.
Generative AI can identify patterns in data, summarize text, personalize data and create content. It can deliver all its results conversationally.
This is disruptive
Sarrazin believes generative AI's conversational ability is disruptive. He compares it to the arrival of HTML and mobile computing. That has implications for what he does.
He has asked his technical team, "If we're going to make [generative AI] a primary interface, how would you design the product differently?"
If we're going to make [generative AI] a primary interface, how would you design the product differently?
Hugo SarrazinChief product and technology officer, UKG
"We're in the process of working across the whole suite and trying to find the right way to create conversational AI," Sarrazin said. He sees its application in recruiting, performance management, payroll, and workforce management.
UKG signaled where it was going with generative AI this month. It said it would use Google's Vertex AI enterprise cloud platform for its generative AI development. It has used Google's AI capabilities since 2016 for predictive analytics and sentiment analysis.
Vertex has the large language model capabilities of Google's consumer product, Bard. But UKG data and customer data are layered on Vertex for specialized applications, Sarrazin said.
Generative AI risks are a potential threat to customer adoption. Risks include errors and training data bias. One example of discrimination is hiring recommendations favoring men over women for specific jobs. Federal agencies warned HR vendors and customers that discrimination in hiring could lead to enforcement actions.
Analysts say that any HR tech vendor with generative AI technology will face customer questions about how it works.
The concerns
Generative AI, for instance, could give employers a headstart in performance reviews and provide context, but HR analysts also see problems ahead.
There are broadly several concerns cited by HR analysts. Those include the accuracy of AI-generated data and other ethical considerations, such as a literal interpretation of information that could give too much or too little weight to the wrong thing.
For HR, the issues have been most evident in the area of recruiting tech. Gartner, in a recent report, said as more oganizations leverage AI in recuiting, the ethical implications of practices become more urgent. The firm said it expects this issue to become a crisis this year, "particularly as governments begin scrutinizing the use of AI in hiring."
Ray Wang, an analyst at Constellation Research, who attended a recent UKG analyst presentation about its generative AI plans, said AI will have a range of impacts across HR. They include increasing operational efficiency, improving regulatory compliance by identifying risks earlier and bettering employee experience by identifying concerns.
To offset the risks, Wang said, ethical AI has to be transparent and "open to inspection for all participants," among other things.
Sarrazin said UKG can address ethical concerns through the processes they create around AI.
In career development, for instance, generative AI can use all the existing information about employees from internal databases and external systems such as LinkedIn. It can use that information to suggest career opportunities for the worker. It also has data about the employer, he said.
Training AI with the correct data will help with bias risk. But another check puts the manager and employee "in the middle" of the process, Sarrazin said. Both can confirm the system's output.
"You have the ability to correct the information and make sure it is as accurate as possible," he said.
Patrick Thibodeau covers HCM and ERP technologies for TechTarget Editorial. He's worked for more than two decades as an enterprise IT reporter.
