Last month, we posted a blog that describes how you should begin your company’s self-service transformation by identifying the call center processes where incremental steps toward automation are most likely to be successful. In this entry, we’ll discuss how to design the self-service now that you have chosen inquiry types to automate. The best place to start is with your existing live agents.
You have taken great care to set both agents and customers up for success as they work through inquiries together. Agents follow certain steps to handle complex but repeatable transactions, which provide key insights in how to model your self-service powered by Artificial Intelligence (AI). At the very least, it is a great place to start when looking for design inspiration. Using that blueprint, think about ways to improve that approach, if possible. The idea is that AI mimics the capabilities of humans, so design the process to allow for that.
For example, if your live agents are required to collect date of birth and order number to authenticate and access data, your AI -powered self-service should do the same. This is true for customer failures as well; if customers do not know a piece of personally identifiable information (PII), agents move on to the next best piece of PII, and self-service ought to do the same.
This design approach goes back to a tip from the first best practice: you don’t have to disrupt or completely change your service to implement AI. In fact, the only changes will be positive ones. AI often handles inquiries much more efficiently than agents can, in some cases cutting customer effort by up to 50%. This helps you meet customer demand and expectations by offering a low effort, efficient service experience that also saves you money.
A final benefit to designing your self-service to mimic agent processes is that agents will not have to be re-trained and customers will not have to re-learn the rules for how to complete a task. Instead, customers will be given an enhanced experience that is familiar to them.
If your company does not offer live agent service or there is no agent process to mimic, tread carefully when wading into self-service and consult industry experts. The process you are envisioning may not be ideal for automation yet. For example, many brands jumped the gun on chatbots, developing knowledge-based bots that were not designed to handle inquiries in the same way agents were. Customers assumed these bots were “all-knowing” and ended up having poor experiences with them.
Mimicking live agent behavior with AI self-service is all about maintaining a consistent and effortless experience. In the next entry, we’ll discuss how this balance should be achieved not just with voice, but across all of your company’s communication channels.
