Today’s challenging times are proving that customer service and customer retention are more important than ever. Imagine the following all too relatable scenario: you just ordered a case of toilet paper from your favorite online retailer. You’re elated as you managed to find it at a normal price, and it’s set to deliver this week, rather than next month. You wait, delivery day comes and goes, and all the while your front porch remains devoid of the precious package.
As panic sets in, your mind automatically envisions worst-case scenarios. Did some porch pirate swoop in and steal it? Will you have to resort to leaves? No! You decide to follow up with the retailer online and are put in contact with a support representative. This wonderful human comes through for you and manages to locate another case of that white gold, overnighting it to you. The next time you have to order toilet paper, you know exactly which retailer will get your business.
In this scenario, whether you realized it or not, you experienced the final stage of an organization’s customer retention strategy. Put simply, customer retention is an organizational initiative to retain as many customers as possible. According to Bain, it can be up to 25x more costly to acquire new customers. Customer retention is by no means a novel concept, but it has evolved in recent years with the advent of AI in day-to-day business operations. There are many retention strategies that range from old school thinking, like personalized targeted offers and loyalty programs, to more modern AI-driven techniques. Let’s focus on one AI-driven approach: case classification.
AI Improves Customer Retention… One “Case” at a Time
The toilet paper example above highlights a crucial part of retention: customer service, done right. On the surface, customer service is actually quite simple: address customer needs in a timely manner. Where it can fail is in the actual execution. As with everything in business, customer service has many moving parts and complex layers that move a million miles per minute. This is exacerbated in times of stress, such as our current environment — when communication breaks down, the moving parts jam up, and customers are denied assistance at just the moment they truly need it. Thus, being able to shave a small amount of time off your customer support process is not a luxury; it is a necessity.
Enter Einstein Case Classification, an intelligent Salesforce capability that leverages a machine learning model to predict values for case fields. As with all machine learning models, Einstein Case Classification is trained on past case data to ensure it provides results most applicable to your organization. When a new case comes in, the algorithm is used to surface recommended values for fields that you specify. Along with this recommendation comes a confidence level value, or the likelihood that recommendation is correct. Einstein Case Classification can be set up to decide whether or not to automate the acceptance of the recommendation based on that likelihood. Or, the option is provided to merely recommend these values to end-users and allow agents to choose.
This configurable approach gives businesses the ability to tailor the Einstein Case Classification experience to their needs, resulting in massive benefits to both the organization and their customers, including:
- Improved customer service, allowing more informed conversations and faster case resolution
- Time savings, enabling support agents to serve more customers, and customers to have their questions answered promptly
- Improved data quality, which improves reporting, laying the foundation for a completely data-driven organization; as well as providing more accurate and personalized information during customer interactions
- Model performance metrics, permitting the fine-tuning of your implementation and ensuring it only improves with time and data.
Einstein Case Classification in Action
Now that we’ve completed a crash course in what case classification is, let’s see it in action! As in our scenario above, upon not receiving the toilet paper, you decided to make an online inquiry to the retailer. Unbeknownst to you, this created a case in the retailer’s internal customer relationship management (CRM) system. Utilizing CRM best practices for case management, customer-provided keywords such as “delivery” and “failed” were used in combination with product details to route your case to the correct support agent.
Consider the Following Hypothetical
During a stressful time like the COVID-19 pandemic, the influx of cases will skyrocket due to online orders, quickly overwhelming a customer support team. Even the most prepared customer service department can fail to correctly triage their queue and address the critical cases.
To assist with this, Einstein Case Classification has been configured to populate the “Priority” field if the recommendation is more than 90% likely to be correct (determined by confidence level). Our example, a missing order, is a relatively common problem that occurs with fulfillment, providing plenty of high-quality training data. Due to this, the model has a confidence level of 93% when assigning a priority of “High,” automatically classifying it as a high priority case.
This high-priority designation bubbles the case up to the top of the agent’s queue, getting eyes on it faster. Upon opening the case, the agent is presented with an Einstein Case Classification recommendation to set the Case Reason field to “failed delivery.” Upon review of the case, the agent accepts the recommendation, opens up a chat with you, and gets a new order sent your way.
Let’s Review What Happened
In the case creation process, your Einstein Case Classification setup simply populated two fields: priority and case reason. From an operational perspective, this appears to be a relatively minimal return on investment for Einstein Case Classification. However, the simple act of reliably setting these two fields resulted in a more streamlined case resolution process, leading to increased customer satisfaction and, ultimately, higher retention rates.
The Importance of Having an Intelligent Experience Now
The business world is evolving, and data-centric approaches continue to rise in popularity and effectiveness. As this happens, the disparity between firms who adopt intelligent solutions like case classification for their customer retention needs and those that do not have only become more pronounced. If your organization hasn’t already begun building an intelligent experience, ask yourself: how long can you afford not to?
Contact us to see how we can help power your data initiatives.
In my last Tableau blog post, I discussed the advantages to using Tableau as a tool during machine learning model development, both for exploratory data analysis
A turbulent business landscape, along with shelter-in-place mandates, have left customer service representatives dealing with a new challenge: working remotely in addition to heightened complexities due