Self-Service in the age of AI
By Colin Taylor
Self-service has been employed for more than 30 years to allow customers to serve themselves. This has been a mixed blessing, while many of us value the time savings versus waiting endlessly in a queue for the answer to a simple question and to gain a single piece of information, we all have experience with people shouting “AGENT” into their phones when self-service fails.
Technology is a tool, and there is a lot of good technology out there that can underpin successful self-service. But be careful: technology initiatives have the habit of expanding from a narrow scope to an ever-increasing span ranging across departments and silos. The fault here does not lie with the technology, but rather with the imperfect beings defining the scope and parameters for the technology.
“Let’s set up a chatbot,” someone says; and everyone is in support of this idea; but what capabilities should the chatbot have is all too often vague or incomplete. Should we just offer FAQ’s? Should we access the knowledge base? Will we perform data-dips for real-time order status and detail? Will we just work in chat or do we want to have a voicebot with the same capabilities on our IVR (Interactive Voice Response) system? How will we escalate to an agent when the bot cannot discern the customers’ intent? What about when the center is closed? Will we then arrive to find hundreds of emails or voicemails representing each unsuccessful escalation? If so, how do we plan for this in our Workforce Management (WFM)?
Successful self-service requires planning, due diligence, an understanding of the problems you are seeking to solve, as well as a consensus on the capabilities and deliverables that will solve these problems. Fuzzy goals will lead to fuzzy outcomes. With proper planning and discipline, you can deliver effective self-service. But what self-service should you consider?
Look at how your customers and prospects interact with you. Is it primarily by phone, web or chat/text? What is the channel mix, do they employ the same channels for the same reasons, intents or contact types? Are there points in your processes where customers routinely ask for updates (when will it ship, when did it ship, when should I receive it, etc.)?
Artificial Intelligence (AI) increasingly powers self-service today and the capabilities have improved exponentially since the days of “speak agent” on the IVR. For the above customer-facing services, you can develop and deliver chatbots and voicebots that understand the customer’s intent, and can complete tasks and provide responses based on that intent. They can kick-off workflows that may submit forms, process applications, and send emails and/or text messages in real-time or based on future triggers in the workflow. This is all possible today and organizations are offering these services to their customers. Of course, it needs to be built and trained, but it is doable, and you deliver these solutions leveraging AI capabilities such as Natural Language Processing (NLP), Natural Language Understanding (NLU) and Robotic Process Automation (RPA).
AI can also help you when customers need to speak with a human being. Think of AI as being the valet or squire to the agent, finding and fetching the information the agent needs to serve the customer. This can be pulling up FAQ’s, or accessing data in the CRM, ERP or other systems. Whatever the agent needs to know to help the customer can, in theory, be located and presented. Of course, once again this requires detailed and specific design, development and training, but is it achievable today. Others are doing it.
All of the AI-enabled self-service enhancements require time, expertise and cost, but when the result is a reduction in agent handled contacts, this call avoidance strategy can deliver significant savings downstream.
To find out more about how Taylor Reach can help your company with implementing AI in your customer experience platforms, CLICK HERE to schedule a free consultation.