This year, Research powerhouse McKinsey released a report titled “Notes from the AI frontier: Applications and value of deep learning”. The report provides a high-level analysis of the current AI and deep learning space— focusing on which applications currently provide business value, and which are still proving their worth.
McKinsey calls out three specific use-cases where AI injects value right now, one of which is customer service. I found the overall report very insightful, I feel there are a few additional points that could be introduced to the topic of how AI and deep learning add value to customer service today.
The Ability of AI Assistants to Handle Complex Customer Interactions
The main focus of Mckinsey’s AI in customer service example is how AI programs can identify when a customer interaction is going awry, so it can then hand it off to a human operator. There is no doubt this type of redirection is a major value in customer service; but it’s not the whole story.
The most exciting examples of AI in today are when a human and AI work “alongside” one another to see an entire customer interaction through. This means instead of handing the conversation off to a human when it gets too complicated— adding an additional touch point to the customer’s experience— the AI utilizes real-time human guidance to see the entire interaction through. This way the AI receives enhanced, structured human guidance, and the customer gets a smooth, efficient experience.
AI virtual assistants employing this type of collaborative model are rapidly advancing their natural language processing thanks to deep learning processes that incorporate the structured human direction. Ultimately this means the systems can handle increasingly complex interactions— everything from adjusting travel itineraries, to transferring sensitive bank information, to establishing utility service accounts.
Delivering the True OmniChannel Promise
Another use case for AI in customer service is its ability to infuse personalization and cross-system data collection to deliver the true “omnichannel” experience. While omnichannel has long been promised in the customer service realm (the ability of customers to carry on interactions across multiples devices), AI has brought that experience to life by providing true context persistence and personalization.
For example— say you’re chatting with your bank’s automated text bot to adjust account information. Running late to your kid’s soccer game, you jump in the car. You still need to finish the task, but don’t want to text and drive. Instead of ending the conversation, getting in the car and calling the bank back at a different time— starting the process from beginning— you instead tell the messenger bot that you want to continue through voice. The automated assistant confirms, and calls you on your cell. The conversation continues uninterrupted.
The AI assistant identifies the number you are calling from and retrieves not only your account information, but also the context from the interaction you just ended— allowing the conversation to continue seamlessly. Without the AI assistant you likely would have had two separate and repetitive conversations. With AI you were able to carry on the same conversation, and the burden of convenience is put on the technology.
Customer Service is a Pioneering Space for AI
McKinsey’s report hits on some important points— AI and deep learning is assisting human operators in a major way by giving them recommendations on how to handle specific interactions as well as work with them to hand off conversations when needed. But the reality is that the most exciting applications of AI in customer service are the ones where AI and humans work collaboratively to complete tasks together, not just hand them off to one another.
The fact that customer service is one of only three use cases McKinsey references where AI is adding value today indicates the significant progress that has been made in the space. Not too long ago customer service was a behind-the-times industry, but today it is a pioneering use case for Artificial Intelligence.