Technology didn’t stand still in the pandemic: tech changes very likely occurred in your workplace while employees were working at home. Coupled with the day-to-day changes of the new hybrid workplace, HR professionals will need to be prepared to wrestle with the impact on many fronts as they try to balance new product integration, introduction, and training with the returning workforce. And the technology changes are significant.
Consider: the large ERP and HCM solution providers report record numbers of “go-lives” among new cloud-based HCM customers, adding a need for new product training for HR staff and employees. Returning and newly hired employees will need to learn not only how the new product features work, but the implications for how they might do their job differently—particularly in newly-orchestrated hybrid environments. Further, L&D professionals will return to the workplace to see a plethora of other new business applications, requiring on-the-job training on software such as that for financial, production and supply chain management.
And the HCM vendors—no slackers they—have been busy. At least two or three new releases of their cloud software will have been delivered over the pandemic year, often, as in the case of Oracle, for example, with hundreds of new features available per release. This constant addition of features requires testing, deployment and often training as they are released throughout the organization. New technology underlies many of these applications, adding blockchain and far more commonly, artificial intelligence (AI)to the technology mix, increasing the need for governance and employee education.
Product or Platform?
HR leaders have been hit with “AI-enabled” or “AI Inside” messages from vendors for several years now, but its applicability in HR and the business as a whole is not limited to point applications. An AI program analyzes data — ordinarily an immense amount of it — then decides or recommends what should happen next to complete a task. The ability of the application to continually “learn” and adapt from that machine learning is a fundamental aspect of any AI program. The technological areas that make up AI — machine learning, natural language processing (NLP), predictive analytics capabilities and robotic process automation (RPA) − are underlying functions that support a great and growing number of HR business processes. In other words, AI functionality is not an individual application to be purchased, it is an intelligent infrastructure that can underlie many business applications—HR systems among them.
AI is prevalent in much of today’s HR software: It is close to ubiquitous, for example, in the recent talent acquisition applications from both stand-alone and broader-based HCM providers. In this instance, the advantages are many: it makes sourcing, screening, and hiring processes easier, more efficient and better able to address D&I goals. The efficiency and speed of an AI-enabled application has positive cost benefits in that the engine never tires or takes a day off. (CareerBuilder, for one, reports that its updated talent acquisition solution has reduced the cost per candidate by as much as 50% since the product rebuild employing AI.) Expediency in sourcing and rapid time to fill both represent a cost savings to an organization, alleviating the expensive proposition of an empty seat.
Because AI technology is in effect recommending, making choices, and predicting, human vigilance remains necessary to ensure the right data is being fed to the engine and that the output makes sense for the particular organization. Dr. Cathy O’Neil, author and data scientist, makes a very compelling point in discussing the use of AI in decision making: “Algorithms are opinions embedded in code.”1 Those opinions are only good when they relate to the requirements of your business.
Bot Proliferation: AI at Work
Robotic process automation (RPA) utilizes bots – think Alexa or Siri, for example — to replicate human actions on what can be time-consuming questions and tasks. The increasingly widespread use of these natural language chatbots, some of which are capable of understanding and responding to both written and oral input, in human capital management programs are becoming du rigor in talent acquisition, in responding to benefits and overall employment questions, and providing on-the-job self-help and training solutions. Today’s users are accustomed to bots as sources of information and direction and are used to interacting in natural language with bots in many on-line and telephone transactions. In HR, their value is noted for all the repetitive questions employees, especially new employees, might ask: what vacation days does the company take off? How do I get a proof of employment letter? What certifications are required for my new job? Here the benefits are readily apparent: The excessive amount of work time lost by employees searching for information is well documented – up to many hours a week – and AI and bots can play a major role in improving productivity.
As another example, consider bots in talent acquisition, as response to the candidate can be immediate, and often personalized. Tools such as AllyO, now part of HireVue, or iCIMS TextRecruit Ari use natural language processing and machine learning to enable dual-conversation live chat and offer a beginning to intelligent text-based discussions—easy and convenient for the candidate. These applications understand people’s inputs (rather than just code) and learn and ultimately improve feedback and predictions for actions based on that learning over time. These conversational interfaces prove engaging for candidates, and getting bot-supplied answers in real time lessens the need for people-intervention. That saves time for the recruiter, the hiring manager, and the candidate him or herself, again a productivity enhancer.
Addressing Organizational Skill Gaps
What keeps HR professionals—and their corporate executives—awake at night? inability to locate and retain employees with the skills for both today and tomorrow. This is not new—it has been the prevailing sleep-depriving worry of executives for decades. And while AI is not the total solution, there are features within AI that actually can help.
Too often, an employee is type-cast into the role he or she is in, leading managers to assume that the employee is limited by the skills exhibited in that role; this confinement is frequently the root of attrition. Given the deep learning that underlies AI systems, the concept of skills adjacency can be addressed. Beyond reporting capabilities, deep learning is more inferential, drawing on unstructured and often unlabeled data. For example, if an employee or a job candidate knows or can do a certain thing, the software can infer the prerequisites of that skill or knowledge and determine the likelihood of the person having or being able to learn a similar but different skill. AI has the ability to infer skills from, for example, prior jobs held, volunteer positions, outside interests or education that are not directly stated on employment forms. Because the program is always learning, it retains associations it can then apply in new and different scenarios. While that may sound magical, it actually can be determined based on analysis of millions of datapoints. The implications within HR are many as choices in upskilling employees and facilitating a more inclusive pool of new hires can be addressed, and increasing the possible positions within an organization for which the employee may be suited.
Vendors Suggest, HR Decides
There are global standards on the ethical use of AI, such as the AIHLEG Ethics Guidelines for Trustworthy AI2 driven by the European Commission and the OECD Principles on AI3 which influenced the AI principles adopted by the G20. Vendors often publish their ethical standards for AI and ML based on these and position many of their AI tools, especially those used in talent acquisition, performance reviewing, and promotion recommendations as “suggestions” rather than predictions or mandates. Ultimately, it is you the HR professional who makes a decision, based on those suggestions.
Conclusion
AI learning systems do just what they are supposed to: They “learn” and apply rules based on data gathered over time. What the application learns affects how it behaves — hence it adapts. While the decisions made based on AI may prove correct for your organization, you still have to make sure that the assumptions applied over time are understood and reviewed for accuracy, fairness and relevance to organizational goals. And they need to be “explainable.” As desirable and useful as implementing AI in HR has proved to be, when HR makes decisions based on computer-derived algorithms, someone in the organization must understand what factors are used in creating the correlation or the prediction and how the algorithms work. Transparency and accountability are critical. As with many aspects of technology, HR are needs to be alert to unintended consequences. Understanding the potential and the limitations of AI, especially in its initial forays into the workplace, is key to its judicious and advantageous use.
