I’ll be honest with you – artificial intelligence (AI) is complicated, and when people talk about AI in HR, they often gloss over that fact. There are tons of specific terms associated with it, processes that go along with it like machine learning, and technologies designed to use it. It’s hard enough for data scientists to keep it all straight, let alone an HR professional with a ton of other competing priorities.
I’m sure leading with that did absolutely nothing to calm your anxieties around this topic, but as usual I promise I’ll turn this around into something good. In that spirit, let me revise my first statement a little – AI is complicated, but it doesn’t have to be complicated for you. Here’s what I mean by that in a few quick examples to set you and your HR team’s minds at ease:
- AI uses a variety of different algorithms to generate results, but you don’t need to know them all.
- AI processes are complex to build and maintain, but you don’t need to build them or maintain them.
- AI can automate many tasks, but it won’t take the place of you or your HR team.
So now that you know what you don’t need to know or worry about, what should you be focusing on when it comes to using AI in HR? Well, it really boils down to two things – knowing the differences between the core terms used around AI so you can get a big-picture view and having some key points you can keep in your back pocket to help your organization understand that AI is an advantage, not a threat. Let’s explore.
What is AI anyway?
There are a bunch of ways to describe what AI is and what it does, but I really like how HR Technologist put it when they said “AI enables machines to ‘think like humans.’” I know that may sound too simplistic, but it’s really what matters most when you think about it from an HR perspective.
If you want to see what I mean more clearly, consider these examples of how AI can function in an HR department:
- Dig through employee activity, recent milestones, feedback, pay, and other categories of information to determine which employees are a flight risk.
- Understand the cycles of HR team activity at an organization and proactively serve up personalized resources, tips, and links in an HCM system ahead of important tasks so they are completed efficiently and effectively.
- Find manager activities or behaviors that don't match what's normal for the organization, alerting HR to these anomalies to prevent noncompliance or fraud.
- Analyze the language being used in performance reviews, peer-to-peer feedback, and comments on goals to predict employee success levels and recommend courses of action.
As you can see, these scenarios all have to do with learning from people data, solving for particular problems, reasoning out what’s likely to happen next or what a good recommendation would be, and understanding the language and numbers available enough to come up with a solution that makes sense. And when you remember that it’s your HR technology doing this, then it becomes clear why “thinking like a human” is not a bad way of describing AI.
Now I know this is already raising red flags about the power dynamic between humans and AI in the workplace – we’re going to get into that more deeply later, so bear with me. For now, take a quick look at the examples I just gave in a little bit of a different light. Instead of thinking about them as AI tasks, imagine instead if any one of the projects above was a request coming down from your executive team. I think it’s safe to say it would be a lot added to your plate, right? What about if you were asked to handle all four? We’ll come back to this but I think you’re starting to see what I’m getting at.
Bottom line, AI is a set of processes to help you gather the insights needed to accomplish complex strategic goals that will push your HR organization forward without adding a ton of effort or workload. So with that baseline definition out of the way, let’s walk through a few more terms in the AI space that will help deepen your understanding of how that can happen.
How is AI different than machine learning?
Something I see a lot in AI conversations is the term “machine learning” being thrown around, sometimes interchangeably with AI itself. That can make things more confusing than they need to be. The simple answer here is that machine learning is one of the main ways AI becomes able to do the human-style thinking we just discussed. It’s a set of processes AI can use to analyze different kinds of data, learn from what it finds, and continue to improve its ability to solve the problem it’s tackling over time.
To put all that in more concrete terms, think about when you’re onboarding a new employee. Usually, you give them a baseline to start from – the key pieces of information they need, processes to train for their responsibilities, and any other reference points like mentors or expert co-workers they should touch base with. After that, it’s up to them to find the patterns of behavior they should follow based on the role they’re performing, look for ways to continually get better at it, and ultimately generate results for the organization.
That’s basically what machine learning tries to do from a technical perspective to feed AI results. These processes apply a specific set of rules to some data based around something like recruiting, let’s say, and then learn from how hiring trends change based on any predictions or recommendations they come up with to continuously get better at addressing the issue being tackled.
So to sum up, if AI is the “what” then machine learning is the “how.” There are several other methods AI uses to arrive at answers as well, but machine learning is the one that most frequently gets brought up in discussion. And even if you encounter other AI methods, remember that the goal will always be the same – deliver results that get better over time and help you come to decisions faster or remove steps in complex processes to speed up your actions.
Wait, what about people analytics? Where does that fit in?
People analytics and AI have an interesting relationship. What helps here is thinking about the two like overlapping circles in a Venn diagram. AI does some things that people analytics doesn’t, like process automation, and people analytics is much broader than just AI since it also covers anything to do with day-to-day reporting, predicting trends, and generating recommendations based on data. But when AI and people analytics come together, especially in HR, the results make a huge difference for organizations.
Think about it – the two priorities at the heart of both AI and people analytics are helping HR make informed decisions faster and elevating HR as a strategic partner to the business. Clearly a natural pairing. So how do they come together around those priorities to help HR succeed?
You guessed it – it’s time for another example. To connect it all together, let’s use the first bullet I mentioned when we talked about AI and see how these areas can come together to proactively address the problem of flight risk.
So as we all probably know by now in the HR space, retention is more critical to the survival of our organizations than ever before. And a big part of making sure we can stay ahead of potential threats to retention is knowing when people are likely to leave our companies, what factors are contributing to that desire to leave, and how we can proactively re-engage those people. AI can help us track down the insights we need to realize those goals and connect with our wider people analytics ecosystem to point us to the right data points fast so we can focus on taking action to prevent turnover instead of having to play catch-up.
Here's how something like that would work:
- AI, powered by one of the machine learning algorithms we talked about in the last section, investigates a wide variety of metrics that could be contributing to employees wanting to leave. Things like recent pay changes, overtime levels, performance ratings, incidents, and missed or late punches – among a host of other factors – would all be fair game as considerations.
- As the AI filters through these factors, it rates their significance and figures out what the biggest contributors are.
- Based on the contributing factors, the machine learning process generates predictions around which employees have the highest risk of leaving the company within the next three months.
- These predictions get served up as alerts to the HR team, connecting with related analytics displays like employee succession plan scores to identify high-priority at-risk employees. Links to the right pieces of data around each of the contributing factors identified are also included, as well as links to recommended actions.
- Using the contributing factor links, HR professionals can dive straight into specific areas of their people data, using the data exploration, reporting, and data visualization tools in their HCM system to verify the AI’s results and set up people analytics displays on those metrics. Then they can take action from the same location too if the recommendations seem to make sense, speeding up the decision-making process.
See how all that comes together? AI speeds up the data exploration and prediction process, and then you as the HR expert get to quickly investigate the most relevant areas, prove out the results, and most importantly take proactive action to keep your strongest employees at your organization longer.
Should I be worried about AI's impact on HR and my organization?
And now we come to the elephant in the room that most people think about but are often afraid to say out loud when they encounter AI – trust. What assurances are there that AI and machine learning processes won’t take over HR jobs or take away decision-making power from HR leaders? How do we know the results we get from AI aren’t biased or putting us at risk for compliance problems? These are both good questions to be asking, and luckily there are some clear answers to be had in response.
Since AI’s inception, people’s jobs being automated away has been an ongoing concern. When you consider using AI in HR, you’ll have to address employee anxiety around this head-on to ensure they continue to trust your HCM systems and adopt the solutions you invest in. Think of making a game plan for this as part of your change management strategy.
What you’ll need to ease the fears you encounter is a combination of facts and benefits. On the fact front, a good example is the McKinsey study cited in this article that found that only 5 percent of US jobs could be automated fully while 60 percent could be partly automated. This leads nicely into the second piece around the benefits of AI – it’s not about replacing people, it’s about making their jobs easier and more impactful.
Ultimately you always need a human on the other end of any AI result to make a decision based off of it and take action. AI just gets a bunch of the administrative process you’d normally have to go through to get to those decisions out of the way. The part of your job that gets automated allows you to focus on being more proactive and strategic, moving you and your organization toward your goals at an accelerated pace.
The second common AI anxiety, bias, is a more complex topic than the first one. Here you have to do your research to ensure you’re using a solution that is using AI and machine learning processes responsibly. There are standards out there in the marketplace designed to prevent bias from occurring – make sure you know those and whether or not your AI solution complies with them. Google’s AI principles are a good example of these kinds of standards. Also, look for AI that prioritizes what people do more than who they are, taking behaviors that can be quantified into consideration more heavily than demographics or other inherent characteristics of employees. If you make sure you do your due diligence in this area before you implement AI in your HR department, you’ll mitigate potential risks and further increase how secure employees feel about the new technology.
Conclusion: Adding AI to your HR strategy opens a world of new possibilities
Even all the way back in 2018, PwC was saying that latecomers to AI would quickly find themselves at a competitive disadvantage. That means the time to act is now if you want your HR approach to stand out from the pack and help you recruit, hire, manage, pay, and retain the best employees to help your organization succeed.
If you want to understand all the places where AI could potentially have an impact at your organization, you need to understand the moments that matter for the different people who work there – including yourself. That way when you roll out an AI solution, you’ll be sure it will have a recognizable positive impact on the many different stakeholders HR serves.