Points of View

AI in HR Is a Matter of How, Not If

Aug 13, 2018 Maria Terekhova

One of the most controversial areas in which artificial intelligence (AI) is gaining traction is human resources (HR). There are use cases for AI across the whole spectrum of HR functions, from recruitment to performance management to termination decisions. The use of AI in this department is particularly risky and sensitive because HR is a fundamentally a soft-skill-based function, and any technology deployed within it can have profoundly adverse impacts on compliance and privacy issues.

 

In addition, most ambitious enterprises are determined to become immersed in digital technologies to drive digital business models and bring customers and employees much closer together. Hence smart management of data, where AI is emerging as a primary conduit, is critical for driving staff interaction and performance.

 

AI is not ready to replace HR professionals; it augments complex data to support smart HR decisions

 

Artificial intelligence, or AI, is an umbrella term that includes the categories of machine learning and deep learning (ML and DL), computer vision, natural language processing (NLP), sentiment analysis, and virtual and cognitive agents, among others. What all of these subcategories have in common is that none of them yet amount to the original definition of AI—a self-aware piece of software—and all of these components are far more effective than humans at crunching vast quantities of data and detecting patterns to support predictions and decision making. While AI may still be a nascent technology, in many aspects it is already far superior to humans at data processing and analytics for pre-defined tasks. However, it is critical to deploy AI to provide the data that enhances the judgment qualities of people management, not to replace the sensitive HR related disciplines it assists.

 

In fact, AI is now so good at deriving complex insights from data that its use is spreading rapidly across business functions and within industries. It has already made inroads into sectors like banking and logistics and within departments including finance and accounting and marketing.

 

There are legitimate concerns that, if left unaddressed, will hollow out the massive benefits that AI has the potential to deliver

 

As such, this adoption hasn’t been without controversy. Critics raise legitimate concerns; historical biases embedded in the data algorithms and “packaged” black-box scenarios can often result in discrimination, and in some cases, the AI starts making decisions according to a logic that doesn’t accurately reflect the nuances of a real workplace environment.

 

Invasion of workers’ privacy is becoming a significant issue as algorithms become capable of tracking ever-more sensitive and granular types of input, such as conversations and keystrokes. These are all legitimate concerns that, if left unaddressed, will hollow out the massive benefits that AI has the potential to deliver.

 

The creation of third-party AI offerings that track HR is accelerating

 

The use of AI in HR is nevertheless on the rise. Corporations including Unilever, Hiscox, Nvidia, and Nasdaq have all implemented AI software to help them manage their workforces, either through acquisitions, building their own platforms, or using third-party offerings. And there are many of the latter—HR AI (for want of a better term) is a booming market, with start-ups everywhere building software to help companies take a more data-driven management approach. Each of these products specializes in tracking and analyzing different metrics, and each one is superbly calibrated to solve a particular HR challenge. But what they all promise to do is help enterprise users boost employee productivity, reduce costs, and help companies stay competitive in an increasingly digital economy (see Exhibit 1).

 

The fact is that it’s now too late for companies to be asking themselves, “Should I be using AI to manage my workforce?” Too many of their peers are already using AI; it’s a necessity for remaining competitive, and they don’t have the liberty of not following suit. Companies that want to survive in a digital economy should be asking themselves, “How should I implement AI in my HR department to mitigate the risks and get the highest ROI possible?”

 

Exhibit 1: Machine learning’s greatest potential benefits

 

Source: HFS Research State of Machine Learning 2018, n=153

 

AI is now well-equipped to create a data-driven work environment—here are multiple examples

 

To understand why companies are taking the risk of implementing AI in a department as reliant on soft skills as HR, we have to look at real use cases and what companies stand to gain from this gamble. For companies across all industries and geographies, cutting costs while boosting productivity and revenue is now a fact of life, and it is a constant struggle in an environment in which more technically agile players with minimal infrastructure costs and overhead keep piling into the race. The winners will be those who can effectively use the vast quantities of data they produce, for example, to identify where they can introduce efficiencies, streamline costs, and respond predictively, not reactively, to problems. This type of improvement in resource usage is where AI is making a noticeable impact. The table below (see Exhibit 2) gives just a few examples of where companies are using AI in their HR processes.

 

Exhibit 2: AI in HR use cases

 

Source: HFS Research, 2018

 

As this table demonstrates, AI has the potential to help employers make their workplaces more efficient, secure, compliant, and safe, both for the company as a whole and for each worker. However, what becomes conspicuous when these use cases are placed side by side is that they all take a highly quantitative approach to gauging performance—all of these tools do no more than check whether employees are reaching predetermined targets or complying with predetermined regulations—even if they do that well. They are simple and thorough measures of whether or not employees are staying above previously set thresholds (see Exhibit 3).

 

On the one hand, this is a bonus for most enterprises, as keeping track of whether such boxes were checked would be very hard to do manually by human HR and compliance teams, especially as corporations generate more data as their businesses digitize. But, on the other hand, these solutions also betray a challenge that all companies face, but which algorithms cannot solve—keeping employees motivated and passionate.

 

Exhibit 3: KnowYourDay features

 

 

Source: HFS Research (screenshot)

 

AI can help ensure your workers are doing the bare minimum—but it can't do more than that

 

A perennial problem faced by most companies is ensuring not only that their employees are doing the bare minimum they are contractually obligated to do, such as reaching a required sales quota per month, but also that employees are so dedicated to and passionate about the company they work for that they will go above and beyond the dry requirements in their work contractbuilding robust and intimate client relationships that ensure customer stickiness and satisfaction, for example. This is when AI creates value for enterprises—when humans use their soft skills and creativity not only to execute but also to improve upon the status quo to help their employers stay competitive in a fast-changing economy.

 

And this is where the problem with HR AI lies—it can only measure employee performance according to a very rigid set of metrics. Even the most sophisticated NLP and sentiment analysis software cannot accurately gauge whether an individual’s values align with their employers’ and whether their work motivates them to go the extra mile for their company. The lesson here is that data only goes so far; the insights AI generates then have to be leveraged in a more holistic way, that is rounded out with human insight and judgment, to add meaningful value to companies.

 

For example, software like StatusToday’s can indicate to an employer which of their workers are star performers, but it can’t identify what drives them to perform so well, and thus how to give them assignments that will continue to stimulate them years into the future. Such software can also pin down workers who aren’t being productive, but it won’t tell an employer how to change that worker’s environment to get them motivated, use the impressive skill set that got them hired in the first place, or boost revenue for the company.

 

How much value AI can bring a company depends on the company's broader HR strategy

 

The central thing to remember is that AI is by no means a silver bullet. It will not fix an undermotivated workforce; it will only help confirm that the workforce is, in fact, undermotivated and unhappy. AI will only really be able to add value for companies that have robust strategies in place to fix such fundamental problems, or ideally to pre-empt them before they happen.

 

Rebecca Butler, Senior Associate at UK law firm Wallace LLP, dismisses notions of AI getting rid of HR staff altogether. For the foreseeable future, it won’t have the soft skills needed for dealing with humans. Instead, she believes, algorithms can free up overworked HR staff to deal with more nuanced tasks and problems in the workplace, by automating data-processing-heavy tasks like application screening, conducting background checks on applicants, and approving holiday requests.

 

In other words, there are several things any company implementing AI in its HR processes must be doing to ensure the technology proves worth the investment, all of which feed into the HFS Digital OneOffice framework that successful businesses will have to adopt (see Exhibit 4). The Digital OneOffice Framework is steeped in work culture becoming more autonomous and inclusive, and manager/employee relationships are becoming more dynamic and fluid (as opposed to the legacy “check the box” meetings). AI tools can provide a terrific “glue” to broaden workflows, speed up interactions, and align employees more tightly to broader business outcomes, such as improved customer experiences.

 

  • Revisit your attitude to FTEs. Insights into how people work should be used not to get rid of FTEs, but to get the most out of them—to motivate them to go beyond the bare minimum and leverage their talents to the utmost. Many companies still take the counterproductive attitude of hiring workers then treating them as costs to be reduced as far as possible rather than regarding them as talents to be nurtured.

  • Recycle and upcycle talent. Talent—especially technical skills and design thinking—will increasingly be at a shortage and fought for as the economy digitizes, so companies need to find a way to retain and make the most of their skilled workers, which will serve them better than simply getting rid of people without investigating why they were underperforming (see Exhibit 5). Butler also adds that such software can lead to better learning and development programs, based on insights into where staff members are struggling and excelling.

  • Keep communications open when it comes to HR AI. The only difference between a surveillance program and an improvement scheme is perception and communication. If a company installs AI software on employees’ computers without full disclosure, it can land in hot water, especially in Europe where GDPR was implemented in May 2018. Implementation must also be presented in the right way—not as ammo to get people fired, but as a tool to help them get the most out of their work. Human managers and HR staff have to be equipped to explain why the employer is collecting data.

  • Ensure you have top-notch managers and executives. Your workers will only be able to perform as well as their managers enable them to. Managers are in charge of assigning tasks to employees and guiding them on their priorities. If you have managers in place who aren’t open to letting individuals pursue tasks that stimulate them and who aren’t willing to listen to employee suggestions and worries, those managers are not going to be able to cultivate and grow their talent; you might need to consider realigning their responsibilities.

  • Enable your HR professionals to do their job, don’t try to replace them. AI does not have intuition or soft skills, the qualities required to have nuanced and often hard and awkward conversations with someone about what they want and what they feel they’re not getting from their job. HR professionals specialize in developing an intuition for handling such situations well and for how to improve dynamics between people and teams. Only human HR professionals are fully equipped to use the data AI gathers on people to help assign them to reassign them to tasks they’d be best at, give them more tasks that motivate them personally, or add automation to eliminate tasks they find valueless.

Exhibit 4: The HFS Digital OneOffice Framework

 

Source: HFS Research

 

Exhibit 5: Staff retention challenges among global companies

 

 

Source: ManpowerGroup, 2018, n=39,195

 

Bottom line: Revitalize your corporate values today if you're considering bringing AI into your HR processes

 

Getting employee consent and understanding will be key to introducing AI into the workplace for the benefit of corporations and the people who work for them. A well-planned presentation is essential, says Butler: the crux is framing it correctly. Don’t present the software as something that will give employers ammunition to get rid of unsatisfactory workers. Instead, introduce it as a tool that is excellent for improving productivity and that will get rid of useless work and create more meaningful roles.

 

The first step is to take a look at corporate values and culture, and management style. If you have management in place that takes a cost-reduction approach to FTEs, their approach must be altered and brought in line with company values. AI is a really good car—but it won’t add any value if a bad driver is using it.

 

In other words, a company that begins with the assumption that its workers’ agendas conflict with its own will sour its employee relationships and see its top talent siphoned away by hungry competitors. It will be the companies that see workers as a most valuable resource to be nurtured and encouraged that will survive at the intersection of the war for talent and the advent of AI.