“Those who cannot remember the past are condemned to repeat it.” -George Santayana
In the early 2000s, or futuristically, at the turn of the century, I had the pleasure of working for an organization that ushered in the manic rush from DTMF/Touchtone systems to Interactive Voice Response Systems using directed dialog speech recognition. At the time, speech recognition was at the height of the technology hype cycle and wouldn’t truly be effective for many years. However, the issue mix that agents had to address changed. It happened slowly, and I see it happening again but this time the change will be more pronounced.
Providing self-service for simple tasks has a profound impact on a contact center. And as self-service technology improves with the advent of Robotic Process Automation, Cognitive Computing, and eventually Machine Learning the issue mix of your contact center will change. If you’ve ever been a contact center agent, you’ll know that having a mix of easy 30 second calls and interesting complex customer issues breaks the monotony of an otherwise difficult job. If we rank the difficulty of issues a customer presents to your organization on a scale of 1 to 10 with 10 being the most complex and 1 being easily addressed, then map that against the effort of implementing automation to resolve those issues. We end up with something like the figure below. The yellow line represents the addressable issues using today’s set of automation tools. Nothing too advanced but a good IVR with decent containment, a well-designed web portal, and a fully functioning mobile app is par for the course. As we look to more advanced technologies such as chatbots, cognitive conversation engines, and ever-improving natural language speech recognition, we see a projected red line creeping up the scale of difficulty.
This is all to say that automation is going to change the issue mix your agents face today, just like speech recognition changed the mix at the turn of the century. However, I believe this change is going to be more drastic. Not because the technology is more effective but because we are rapidly approaching a Pareto Distribution and this has serious implications.
- Average Handle Time will drift up because the easily answered issues that impact the lower end of the average are now automated and not part of the AHT calculation.
- When agents are only handling issues more complex than a 6, a robust knowledge base becomes a requirement.
- The ratio of supervisors to agents is likely insufficient given the AHT increase and issue mix.
- Utilization targets, in today’s terms, include some softball problems to solve but with automation, a utilization target of 70% will take more energy.
- Employee engagement is the Achilles’ heel to customer service so hiring innate problem solvers and empowering them to solve problems becomes more critical than ever.
Automation is coming. The cost to implement will continue to drop, and the issue mix that your agents face will considerably skew towards the most challenging problem sets. We know this to be true because it happened in contact centers when speech recognition was implemented and automation rates truly began to rise. Let’s work together and find a starting place so that we can drive towards the destination.