The Best of AI: Predicting Agent Burnout and Attrition

blog feature image
Published:  September 29, 2023
Categories:

The best solution to any problem is to prevent it from happening in the first place.

The problem of agent attrition has plagued contact centers for years, but resignations usually come without warning. Until now, center leaders could only react after the fact, and that tactical limitation has always come at a high price.

Agent attrition varies from one customer service operation to another, but a range of 40-70% per year is generally accepted. The cost of replacing an agent also varies by circumstance, but again, it’s generally agreed that recruiting, hiring, onboarding, and training a replacement agent can cost as much as $35,000. And that doesn’t include the unquantifiable but very real productivity benefits that come with time on the job.

Let’s say an operation employs 1,000 agents and loses 40% of them to attrition each year. That’s 400 agents at $35,000 apiece—or $14 million per year in replacement costs.

Work-related stress among contact center agents is unavoidable, of course. But cumulative stress—the kind that fuels attrition—is preventable. More than three-fourths of agents report high stress driven by time pressure, demanding customers, and complex administrative processes. Stress accumulates quickly and when it’s not addressed, it can lead to burnout and eventually, attrition.

That’s how it’s always been. But things may soon be different, thanks to a new tool that offers contact center leaders a proactive option to address on-the-job agent stress and attrition.

Intradiem, the leading provider of contact center automation solutions, has introduced the first AI-powered technology solution to quantify agent burnout and predict the risk of eventual attrition. The solution currently delivers 80% accuracy, and refinements and model training now in progress are expected to drive that rate past 90% in 2024.

Intradiem has harnessed AI’s predictive power to identify patterns that indicate approaching burnout, such as more frequent absences and lower productivity. Drawing and connecting insights from a broad range of contact center systems’ data, the solution creates a snapshot of each individual agent’s burnout risk. Critically, hourly and longer-term performance data are blended to account for seasonality and other expected variations. The solution assigns each agent to a burnout risk category (low, moderate, high, or critical), which helps supervisors prioritize remedial actions to support agents who are most at risk.

Data indicators are updated hourly and agent scores are reaggregated daily. Supervisors can access these insights at a glance in an intuitive and highly visual dashboard. The solution recommends support actions, ranging from one-to-one meetings with supervisors, to schedule and work queue changes, specific training, additional breaks, and others. Filtering options also allow supervisors to investigate attrition risk at various levels of granularity—from individual agents to managers, groups, or business units—so they can quickly assess if specific teams or managers are associated with higher attrition rates.

In recent years, customer service performance has taken on greater significance in the battle to maintain a positive brand reputation at large organizations. Customer-facing agents are on the front lines of that battle, and keeping them engaged, satisfied, and productive requires constant and detailed attention. But contact center supervisors’ attention is already spread thinly across multiple responsibilities, and few have the bandwidth to dedicate to the state of mind of individual agents.

This is a perfect application area for powerful new AI capabilities. Intradiem has seized the opportunity to create a proactive new tool that allows supervisors to foresee burnout and forestall attrition. Armed in advance with critical insights, they can prioritize responses and keep their teams together. That will save a lot of money, a lot of time, and a lot of great agents.

Categories

Archive