In the world of Customer Service, Artificial Intelligence (AI) is one of the hottest buzzwords of 2019, followed closely by Robotic Process Automation (RPA). True AI utilizes “models” that learn and adapt to customers, self-optimizing over time. Moving to AI is challenging not only from the technical perspective, but also from the business perspective as the benefits are often not quantifiable and have yet to be proven. While AI software can identify relevancy and correlation of factors a human analyst might miss, starting with an analysis-driven approach can build confidence, show ROI, and lay the foundation for a more autonomous future.
In the Customer Service space, AI can be used to identify customer situations and determine how to best handle them, including directing them to the best agent or function to support their situation. Focusing on a few initial use cases can make the process more transparent, show ROI, and generate buy-in. Let’s take the case of “retention” as an initial example. Identifying key customer attributes that can predict that a customer contact is retention-driven can improve customer service operations (less agent transfers) and business metrics (better retention rates). (Take this one step further in the future, proactively engaging potential churn customers before they even get a chance to leave.)
Start With a Core Use Case and Human Analysis
Start with some basic data analytics and examine key customer attributes, including customer contacts, and the contact results. As an example, look at how many days from the contact until the customer’s subscription expires and compare that to call reason. You may find that if a customer contacts you within X days of their expiration, then Y% of the time it is regarding cancelling or renegotiating. There may be one or two other key factors to examine. Keep the list small, you can expand later. The more complicated the analysis rules, the harder it will be to get others on board.
Use the Data and Results to Build the Business Cases and Confidence
Use the data to make the business case. If those contacts were now automatically routed to Retention agents, how much would be saved in agent time and/or improved in retention rate? This is the business case that starts to show the value of customer information driven handling. And implementation will prove the results.
Build a Foundational Architecture for Future Complexities
Implementation architecture also becomes key to laying a future foundation. Separate the analysis engine from the channels. Modify the channels to reach out to the analysis component, allowing the analysis engine to make the recommendation to the channel. The channel just needs to know how to implement the recommendation, not how to determine it.
As you implement a few use cases, you will be gaining acceptance that attribute-driven customer service is a benefit to your organization. You can then start expanding the models and start looking at more automated, self-learning models and tools. Because you have architected the analysis engine component as a separate engine, it can become more and more sophisticated with minimal impact to the individual customer service channel technologies. This approach not only builds the capabilities iteratively, it also builds the business buy-in.