Is Your Enterprise Chatbot-Ready?

If the explosion of AI-powered technologies over the last few years has passed you by, then you may have missed the rise of the chatbot. According to market research firm Marketsandmarkets, the chatbot market reached $703 million in 2016, and is estimated to grow with a compounded annual growth rate (CAGR) of 35.2 percent between 2016 and 2022. More advanced technology and improved deployments are indeed bringing chatbot technologies to the front line of customer engagement. But at what point do you know a chatbot will add value to your organization? How do you know when your business is “chatbot ready?”

 

Chatbots have a number of current uses – from engaging with customers to build a brand, to placing a pizza order or even discussing your finances – but the area of greatest interest for enterprise buyers at present is cost savings. Most companies look to chatbots as a means to reduce the need for humans and, therefore, reduce operational costs. Of course, cost savings isn’t the only benefit of a chatbot as we’ve seen with other automation technologies like RPA. The improved accuracy, speed and increased productivity a chatbot brings have all added to the customer experience. The real promise of a chatbot is its 24/7 availability – no lunch breaks or time off – so a customer has constant access to services.

 

Of course, it’s possible to overinvest in your manual support center. UK fashion powerhouse Burberry, for example, offers 24/7 customer support in 14 languages. But the question stands: do you need a mountain of bilingual fashion experts working nightshifts in contact centers around the world to answer the burning questions of Burberry’s millennial fan base? Or could 80 percent of these questions be answered by a digital employee at a fraction of the cost?

 

To answer that question for Burberry – and for yourself – you first need to understand why your customers are making contact with you. Regardless of whether contact is coming from internal or external customers, the fundamental question remains the same. How many inbound contacts do you receive? What are the reasons for contact? And how complex are they to resolve?

 

It is easy to get caught up in the complex nature of the queries – how everyone is different and how you want to connect with your customers in personalized ways – but getting into this kind of detail too early will slow you down. The way to begin and quickly gain traction is this:  Start with the numbers and assess your total opportunity. Figure out what portion of your volume can be bucketed into 20 or 30 defined customer journeys. This becomes your targeted opportunity.

 

Keep in mind that you won’t be able to automate everything, at least initially. Trying to create a great experience for some can result in a poor experience for everyone. And don’t forget that a chatbot is unlikely to be the solution for every query – or every category of query. A smooth handover to an actual person for particular situations needs to be built into your business case.

 

Once you understand your targeted opportunity, it’s time to get into the detail. Get a handle on the processes these contacts are following. What is the input or intent of the customer? How is this conveyed? What is the process the agent follows? What is the outcome for the customer?

 

Each of these components should be scored against a complexity matrix, which, once completed for all your chosen user journeys, will give you another metric to assess your chatbot readiness. This is called your “complexity to automate.” In this context, complexity takes on two meanings: the complexity of the conversation and the complexity of the resolution. Don’t be surprised that simple resolution processes can stem from complex conversations and vice versa. Even on a case-by-case basis, complexity can change. These are important factors to include in your initial high-level business case.

 

This brings us to the process of selecting the right technology, which presents its own set of challenges and degrees of complexity. Not only are there many vendors to choose from, but different products are capable of different capabilities. Will a Q&A-type chatbot technology fit the bill? Or does your chatbot need to be capable of a conversation? Assessing your operation puts you in a better place to home in on the right technology.

 

Unfortunately, the market is full of hype at the moment, so we recommend double checking client references and trying a “hack-a-thon style” procurement process to see, touch and learn first-hand about how each product will meet your needs in a face-to-face, hands-on environment. There is no better way to vet capability than by seeing a live demonstration!