AI and Machine Learning (ML) are all the rage these days, and rightly so. Like the Internet 20 years ago, AI/ML is a pervasive change that will affect nearly every function of every business. AI/ML are changing business workflows in sales, customer service, marketing, finance, strategic planning to mention a few of the functions. HR is no exception.
HR leaders need to develop a comprehensive AI strategy to prepare for a work environment where AI is omnipresent and co-existent with humans. Specifically, HR needs to utilize and prepare for AI in the following two broad ways:
1. Define strategy to utilize AI within HR, i.e., to:
a. Optimize delivery of HR services
b. Deliver better outcomes
c. Improve talent acquisition
2. Manage AI’s impact on the rest of the enterprise
Optimize Delivery of HR Services
Repetitive tasks like new employee onboarding, benefits questions, vacation, employee training etc. are perfect candidates for being replaced by an AI-based automation solution. Companies that have automated such tasks are seeing productivity and satisfaction gains. According to Debbie Shotwell, Chief People Officer for Saba Software, “automating repeatable workflows has allowed my team to focus on higher-value tasks, leading to increased satisfaction.” Additionally, AI-based systems support a variety of interfaces: text, voice, chat or web, giving employees access via the medium of their choice. This is a low hanging fruit that all HR organizations should consider adopting.
Deliver Better Outcomes
AI is not just limited to automating basic workflows. AI/ML solutions are adept at analyzing troves of unstructured data to provide non-obvious insights and recommendations that can help the HR team deliver outcomes that are not possible today. Some examples are:
• Predict turnover: Imagine if you could predict which of your top employees are considering leaving for greener pastures, so you could try to retain them? In a recent analysis of 130 million employee emails (just metadata, not email content) by McKesson, it was found that high turnover teams had different relationship patterns than low turnover teams. Microsoft also found that analysis of unstructured data like email, slack chats, calendar etc. reveals insights about how well teams are gelling together, which can predict turnover. IBM claims that, Watson, its AI solution, can predict – with a 95% accuracy – if an employee is likely to leave!
• Boosting team productivity: Sometimes a team of individually good, not exceptional, members may outperform a team of individual All-Stars, because they have better chemistry. Google’s project Aristotle utilized extensive data analytics over a period of three years to figure out why some teams work better than others. Joe Militello, Chief People Officer of Pivotal, a strong believer in HR data analytics who employs a People Data Scientist, says, “we’ve looked at findings from Project Aristotle and other analytics to figure out how to boost team productivity”. Currently, this requires manual data analytics, but expect AI/ML solutions to emerge that can recommend ideal team composition based on insights from unstructured data within the company’s communication systems.
"CHROs need to be proactive and lead the way as the enterprise undergoes a transformation that will affect nearly every employee, every business function and the org structure itself"
• Career coach: Employees feel valued when companies are vested in their professional advancement. But most companies don’t have the manpower to provide 1x1 coaching or career planning. An AI based system can provide personalized advice to employees in a structured manner. It can analyze an employee’s current skillset, identify areas of growth and map out possible career progressions. It can also recommend trainings that an employee can take to help them expand their skillset.
• Better performance review and management: The traditional annual performance review is broken. It’s usually out of date, doesn’t reflect “behind-the-scenes” soft skills like collaboration, and can suffer from manager bias, both subconscious and recency bias. There is a move towards a more agile and data-based system, which focuses on ongoing feedback and more reliance on data from multiple sources. AI based systems can help in multiple ways. One, they can gather data from multiple sources: e.g. emails, slack chats, surveys etc. to provide a measure of the “soft” skills. They can also provide an objective assessment of the employee’s skills and impact. The idea is not to replace human judgement, but rather augment the manager’s “gut” feel with a more structured assessment to create a better performance measurement and feedback system.
Clearly, there are privacy implications, especially when AI systems analyze employee communication. While, US laws permit employers to monitor corporate owned devices and networks, there is potential for overreach. HR, as the steward of the employee population, should lead the way in defining ground rules for AI that balance business interests with protecting employee privacy. For example, in the Microsoft example above, Managers are only allowed to view aggregate group data.
Additionally, it is important to remember the human touch. AI and data analytics are nowhere close to replacing humans. Instead, their goal is to augment human decisions. According to Militello, “It is important to remember the human touch as employees are our competitive advantage.”
Improve Talent Acquisition
There has been a lot of focus on AI in recruitment because of the potential to show immediate efficacy and ROI. Some examples are:
• Reduce bias in hiring: Sometimes job descriptions can unwittingly include verbiage that conveys bias, turning off potential women and minority applicants. AI solutions like Textio can analyze job descriptions and suggest changes to make them more gender neutral, thus attracting a more diverse candidate pool.
• Interview scheduling: Interview scheduling can be a tedious and time-consuming task for your recruiting team. There are a number of AI-based solutions available e.g. My Ally, Xor.ai, which can handle complex, multi-party scheduling and integrate with your ATS to automate scheduling, pipeline management and post-interview feedback.
• Augment human decision making about “fit”: Soft skills like communication skills, ability to handle stress and persuasiveness are key indicators of success in addition to the “hard” skills. Using video interviews and then analyzing them via AI, companies can get insights into the candidates’ personality and more accurately predict a fit with the organization culture.
Manage impact of Enterprise wide AI adoption
As every major business function adopts AI, most jobs and even the org structure will undergo a major shift. Employees are concerned: will AI take my job? The truth is that AI is a net creator of jobs. However, several job types will become extinct, and nearly every job will evolve to become AI augmented. Employees performing AI augmented jobs will be significantly more productive, insightful and effective, leading to less bias, more cross-functional collaboration and a flatter org structure.
This is a huge change. Companies that proactively manage, indeed leverage, this change will thrive. Those that resist will likely become extinct. HR has a unique opportunity to lead the way and shepherd this change. As part of strategic workforce planning, HR leaders should identify how AI will change the workforce composition and put a plan in place that identifies needs for retraining, retooling and redeployment of the affected employees. It is also important to communicate with your employees. According to Shotwell, “Explain the rationale to your employees and how the change helps them be more productive and strategic, then guide them through the process to help them adapt.”
Like it or not, AI is the future. According to a Kronos survey 88% of Gen Z employees feel that AI can improve their job in some manner. CHROs need to be proactive and lead the way as the enterprise undergoes a transformation that will affect nearly every employee, every business function and the org structure itself.