Businessolver Blog

Q&A with an Expert: What’s Next for HR and Artificial Intelligence?

Q&A with an Expert: What’s Next for HR and Artificial Intelligence?
Posted on Wednesday, November 20, 2019 by Rae Shanahan
Share:

If there’s one tech trend we’re following this year here at Businessolver, it’s the amazing growth of artificial intelligence in the HR space.

AI-designed-for-HR

AI’s application to human resources has been well-documented, but the exponential growth of this field over the past few years requires HR leaders to keep a watchful eye on where it’s heading.

We spoke to Sony SungChu, Businessolver’s VP of Data Science & Innovation, to understand what’s coming for AI in HR:

What is new in AI designed for HR? How are we moving beyond the chatbot?

There are two primary areas where we see AI going for HR:

  1. System integration. The first is integrating systems across multiple functions such as analytics, recruiting, human capital management, benefits, training, performance management, policy and so on. In a few years, HR teams will be able to consult with one virtual assistant to get information, perform analytical tasks, manage the lifecycle of employees and automate tasks. This will be possible by advances in the technology’s ability to take information and make it useful. For example, AI could “read” HR handbooks and derive the possible questions employees could ask as well as generate the answers to those questions.
  2. Empowering employees. The second area is empowering employees by protecting their identity and keeping their information confidential. AI could elicit or analyze information regarding their experience at the company, their interaction with managers and coworkers, mine insights from the information and provide aggregated actionable data to HR. This might range from recommending an employee for a promotion based on task completion, to offering employees an outlet to freely report harassment and routing the information to HR with a recommended action based on company policy.

In 2020 at Businessolver, our virtual assistant, Sofia, will be breaking new ground by learning topics outside of HR and benefits. This will give her a wide range of knowledge to make inferences when interacting with end-users. She will also have new tools such as analytic capabilities where a user can ask her to analyze data she is given access to.

Have security issues changed recently? If so, how?

Information security remains a hot topic for our industry, and of course AI used for HR must abide by HIPAA and other standards. Safeguarding user information, whether it’s demographic data, utilization, medical, financial or otherwise will continue to remain the number one priority.

One area that will continue to evolve is the ability of AI services to offer choices about sharing information and letting people know how that information is being used. With Sofia’s next generation of features, we look to help users make sense of the available data and only utilize it when they give permission to do so.

An example would be someone who asks Sofia to analyze their day-to-day expenses and project their ability to cover a medical emergency. If the user gives Sofia permission to access financial data, it will never be shared and will be used only within the context agreed upon by the user.

What is the low-hanging fruit with AI implementation for HR?

The easiest implementation of AI is a virtual assistant that can retrieve information for employees or confirm the status of transactions. For example, “Did my PTO request get approved?” “How do I submit my expense reports?” “Did I hit my 401(k) contribution limit?” AI can also help elicit information from employees via surveys or process forms.

Which areas of HR function or technology are challenging for AI solutions implementation?

It’s a challenge to make clear that AI is still artificial intelligence, meaning you can still “trip it up.” Even with the most sophisticated AI today, it’s still fairly easy to come across utterances or circumstances in interactions where the AI gets it completely wrong or can’t help you—such as asking Alexa to play a song with a Spanish title, or asking Sofia to call a taxi for you. That said, there is something very human about the way machines learn. Think of a person learning a new skill or a new language. It takes time to get good at something. It’s the same with AI.

Do we have data on ROI now that AI solutions are entering the landscape?

Absolutely. Our own application illustrates how useful it’s been. For the most part, you’ll hear about cutting costs of HR teams and freeing up time, so HR teams don’t have to work on mundane tasks. The biggest ROI coming in the following years will center around what AI does for users. How is AI helping people make better decisions or get more out of the sea of data that surrounds us? Even in the HR space, it’s easier to show up at our website and ask Sofia about dependent verification than sifting through documents or calling member services. 

What trends are on the horizon for 2020?

As AI becomes more accessible, consumers will need to be a little more savvy about their choices. More and more companies will offer their own flavor on AI and promise quick configurability, ease of use and speedy responses. As consumers, we’ll need to understand the edge cases. For example, if I talk to the service, can the AI assistant understand an accent and still interpret my intent correctly? Will I need to feed it additional data to train it, and if so, how much? Is it secure?

Specific to HR, I can see a wider adoption of AI for analytics. I see AI being used to help in the Employee Assistance Program space where anonymity is important. I can also see AI playing a bigger role in mobile devices, allowing people to grant AI access to a myriad of applications and, in particular, helping people make decisions by recommending actions or helping them access more useful resources.

But when developing AI you also need to think about empathy. Can machines demonstrate empathy when explaining benefits information? Find out. 

{{cta(‘1be86c94-dfd3-41c7-bdf3-8e420cd3e949’)}}