AI Can Help with Employee Engagement and Retention – But What Are the Risks and Best Practices?

A recent Gallup poll shows that only 31% of employees feel actively engaged at work, with the majority of workers voicing concerns about whether their supervisor cares about them as a person, their lack of opportunity to learn and grow, and not understanding their company’s mission or purpose. Could AI technology hold the key to revival? This Insight will outline some of the ways that AI could assist your organization in connecting with your workforce, point out the risks, and then provide a playbook for how AI can aid in your employee engagement and retention efforts.

AI Offers a Predictive Analytics Solution

By using AI to leverage data analytics, employers can proactively identify the factors that contribute to resignations and quiet-quitting and implement strategies to retain talent. Some commonly analyzed factors include:

• Pay increases: Analytics can show whether employees who receive regular pay increases are less likely to resign compared to stagnant compensation.
• Training opportunities: The ability to excel in a role with access to training can lead to higher retention. Data can show what training is helpful and if employee training leads to employees staying with the company.
• Promotions: Waiting for promotion or advancement can be a significant driver of turnover. Predictive models can help quantify the impact of delayed promotions on resignation rates.
• Benefits: The benefits structure can be analyzed to determine its effect on employee retention.
• Changes to management structure: Organizational changes, such as new management or restructuring, can disrupt employee engagement. Analytics can help assess whether such changes are linked to increased resignations.
• Employee behavior: By reviewing email and other communications, attendance at meetings, and general work activity, programs can assist in determining how engaged your workers are.

All these factors can be analyzed together using AI to determine the risk of an employee leaving. This allows your organization to proactively engage with that employee and address underlying issues.

What Are the Risks When Relying on AI For Employee Engagement Analysis?

There are inherent risks when using any type of technology to help track human behavior, and deploying AI to assist your engagement and retention strategy is no different. What are the key risks you should look for?

• False positives. For example, monitoring employees’ attendance at meetings and trainings is not always straightforward. An employee could be not attending meetings because they are busy with large projects or distracted by personal issues unrelated to work. Relying solely on the AI analysis could miss these types of nuance.
• Bias. When analyzing employee retention analytics, models training on historical data can reflect and amplify workplace inequities. This may lead to certain groups being labeled as flight risks.

• Employee privacy concerns. For AI to provide helpful analytics, significant amounts of employee data must be analyzed, which raises privacy concerns. These systems may process data (such as sensitive health data) when determining patterns of absences and other information about employee behavior. Without clear notice and governance, organizations risk losing employee trust.
• Security concerns. Similarly to privacy concerns, the aggregation of the information can create heightened security risks as the vendors providing these tools can become attractive targets for cybercrimes. Further, the use of third parties introduces vulnerabilities, making robust vendor due diligence and contractual security controls important.
• Over-reliance on automated decisions. Once these tools are implemented, it is easy to become over-reliant on the analytics as they are much easier to interpret than reviewing employee records (and often “easier” than direct human communication). Without human oversight, automated predictions may drive certain personnel decisions that overlook nuances.
• Lawsuits. While most lawsuits in the AI workplace arena have related to the hiring process instead of the firing process, utilizing AI to make firing decisions without human oversight could lead to complaints.
• Regulatory concerns. States are beginning to pass legislation about the use of AI to make any employment-related decisions. Using AI analytics to determine employee engagement and potentially use the data to make firing decisions could run you afoul with these laws. See our recent Insight about California bills in this area.

7 Mitigation Steps Your Organization Can Take

Your organization can mitigate the risks of using AI for employee retention analytics by implementing an AI governance program to address common issues. The seven specific recommended steps include:

1. Engage in pre-deployment and ongoing bias testing to identify and remediate issues.
2. Review the data collected to make sure that the models are relying on good data and audit your vendors to ensure that they are reviewing their models.
3. Implement processes for reviewing vendors to ensure that vendors systems align with your company’s expectations (here are the essential questions you should ask your AI vendor).
4. Strengthen your privacy and security measures to keep data safe.
5. Train users on the risks of the automated decisions and provide review guidance.
6. Ensure that human review remains central to any employment-related decisions.
7. Revise your employee handbook and policies to provide clear notice that your organization uses AI analytics and that AI has access to certain employee data.

If you are not sure where to begin, our AI Governance 101 Guide provides a helpful starting point and can be found here.

Conclusion

We will continue to monitor AI and related developments and provide the most up-to-date information directly to your inbox, so make sure you are subscribed to Fisher Phillips’ Insight System. If you have questions, contact your Fisher Phillips attorney, the authors of this Insight, or any attorney in our AI, Data, and Analytics Practice Group.