Diversity, equity, inclusion and belonging (DEIB) have become essential for employers looking to retain workers, increase productivity and stand out in a crowded marketplace. And recent studies show DEIB creates real results for organizations. Gartner found that diverse teams boost team performance by as much as 30 percent. McKinsey & Company found that companies ranking in the highest quartile for diversity are 36 percent more profitable than those in the lowest. And, according to the Harvard Business Review, feeling accepted and valued at work is connected to a 56 percent increase in job performance, 50 percent decline in turnover and 75 percent reduction in sick time.
That’s an annual savings of at least $52 million for a company with 10,000 employees. Most organizations are eager to enjoy these benefits, and most organizations have had at least some efforts to improve DEIB in place for decades—but progress has been painfully slow. One reason may be an under-reliance on analytical and technological approaches.
People decisions are one of the most important, but also one of the most difficult, aspects to measure in an organization. Luckily people analytics—the practice of measuring, collecting, analyzing and transforming data about people and work—has been on a meteoric rise in the last decade. In fact, LinkedIn named the human resources analytics manager, also known as people analytics manager, the second fastest growing job in the U.S. over the last five years. Combining data-informed people decision-making with DEIB goals can give HR and business leaders the edge they need to improve workplace fairness. I’ve even written about using people analytics to improve fairness. But now there’s a way to supercharge those efforts while leveraging the hottest technology trend in decades.
The power of AI
Yes, I’m talking about artificial intelligence. People analytics is all about delivering better talent acquisition, workplace processes and worker experiences through data-driven insights. AI is a natural tool for use in that process. In fact, AI-based people analytics are already in use to predict employee attrition, identify hard-to-find talent, streamline HR operational processes, design training, facilitate onboarding and more. Now AI can be used to advance DEIB efforts by reducing bias in recruiting, hiring, promotion decisions and developmental opportunities. But only if you use it for those purposes.
As humans, we have nearly 200 types of cognitive bias that affect our decision-making. Some of those biases make their way into our decisions—like who to hire and promote and how to structure work – whether we’re aware of them or not. Those decisions then get codified into the data used by AI to make predictions and business decisions. If humans unknowingly introduce bias and then let AI run loose, it could perpetuate those patterns. So, what should an organization looking to leverage data-informed solutions for DEIB do? Combine artificial intelligence with human intelligence.
People analytics powered by AI in the hands of seasoned HR professionals could create a sea change in employment practices. HR professionals understand how to connect the DEIB goals of the organization and the well-being of the workforce through processes and programs. But many HR professionals chose their career to make a difference in people’s lives, not to run analyses, resulting in a lack of data-informed DEIB solutions. Today, with advances in easy-to-use AI tools, HR can focus on what they do best—connecting the business side with the people side—and leverage AI as their analytical assistant.
Almost any organization can use people analytics combined with AI to create a more diverse, inclusive and fair workplace. Here are five approaches to try.
ID patterns
All organizations have data—it is gathered and maintained to support hiring, promotion, pay decisions, bonuses and more. But not all organizations use that data and even fewer use it to progress DEIB efforts. With AI, organizations can go further.
For example, a Harvard Business School study revealed that 75 percent of American employers use some sort of recruitment software, but most limit it to scanning incoming resumes for desired keywords. The remaining insights available in that data go unused, mostly because people don’t know how to mine qualitative data like words on a resume. But AI’s ability to handle large amounts of unstructured data makes it the perfect tool to find patterns in data that doesn’t fit neatly into spreadsheets.
While it may be difficult to identify bias in a single hiring decision, patterns will emerge over multiple instances. With AI, organizations can analyze resumes, identify similarities and differences between individuals hired by the company versus those who were not. Each pattern can then be assessed for how it impacts broader DEIB goals of the organization and modifications to the selection process could be implemented.
Stop biases in their tracks
Reports from Harvard and Indeed have pointed out AI’s potential for bias. But any technology will only be as fair as it is programmed to be. Automated processes can just as easily be programmed to remove bias.
As an example, the natural language processing powers of AI could allow an organization to assess job postings for gender, age or other group-based language biases and remove them. Imagine a recruiting system set to automatically review a hiring manager’s new job listing and screen out unconsciously biased language before it ever gets released onto a job posting site. This approach would create a barrier that blocks unconscious biases before they ever have a chance to enter the workforce decision-making processes.
Program for inclusion
I often tell people: if your organization is perfect, then build a predictive model with your existing data and just follow what it tells you.
I have yet to meet this perfect organization.
Using past data to make decisions about the future will only keep you where you are. If your organization is not meeting its DEIB goals, copying the past won’t get you there. And implementing automated systems based on the patterns that got you there could even set you back further.
Consider as an example an organization that uses a nomination-based promotion process and sees diversity imbalances in promotion rates. Rather than embrace AI for automating the current process, they could use people analytics and AI to more objectively identify individuals for promotion consideration that would not have otherwise been nominated. Placing them into the consideration process may uncover unnoticed and underappreciated superstars in the organization.
There’s no reason this organization would have to stop at simply identifying more diverse individuals for promotion. It could even help create the circumstances that lead to promotion. For example, certain methods used in people analytics, such as organizational network analysis, are well-suited to identifying the interpersonal connections and work opportunities most associated with advancement in an organization. Once identified, HR can create programs to ensure more diverse individuals have access to these opportunities.
Continuously improve
Implementing change is an ongoing process and organizations will continue to generate people data. By feeding that data into AI for analysis, tweaking the solutions to remove biases and monitoring for the introduction of unintended consequences, progress becomes a virtuous cycle. People analytics approaches can be built into AI to uncover patterns of bias. HR teams then develop new unbiased processes. AI looks for signs of improvement while HR continues to adjust as needed, and so on.
For example, many companies assess pay equity, but they do so as a single point-in-time analysis. An improvement-minded, AI-powered organization could design a system to continuously monitor pay equity, incorporating every new data point that occurs as employees flow into, out of and through the organization. AI could even be programmed to consider additional factors and organizational values to create pay scales that can adjust for equity in ways that standard pay scale surveys don’t.
Build in belonging, wellness and inclusion
AI can help simplify the once complicated process of measuring work patterns and interactions within an organization. Many are surprised to learn that it is possible, at least in part, to measure connection and inclusion at work. AI driven people analytics can make quick work of identifying isolated employees—those who are overloaded and at risk for burnout. They can even identify quiet influencers and untapped potential. Once identified, belonging and wellness efforts can be more effectively targeted for the greatest impact on employee engagement and retention.
When seeking to try out a new AI application, remember privacy and trust. Privacy and trust are major concerns when it comes to using AI for anything, even more so when it involves the kind of data collected for people analytics. Organizations need to be cautious—no AI tool should be used for people analytics that hasn’t been approved by someone knowledgeable on the potential impacts. Luckily new regulations and standards are emerging to help guide leaders on this journey. Make friends with your data privacy officer.
Supercharging DEIB efforts is possible through people analytics and AI. AI can uncover areas for improvement and automate some inclusion into practice, but don’t forget that it takes humans to create change. Human intelligence boosted by the power of AI could help to make any workplace a better place.