Chatbots are the hype of 2016 – and there is no sign of it subsiding as we enter 2017. Unfortunately, many bot implementations don’t live up to the inflated expectations. However, chatbots have real value when properly conceived, and one area with tremendous potential is customer service.
Let’s look at some facts:
Fact 1: People hate to repeat themselves, hate to wait on hold — or just outright hate phoning businesses in the first place (2016 Aspect Consumer Experience Index).
By taking the conversation to a messaging channel (to complement IVR), you immediately benefit from an “eternal thread” between you and the business, which you can build upon with every new conversation. Furthermore, you benefit from the asynchronous nature of the channel, so you can pace the dialog at your convenience without any perceived “on hold” times. You can also multi-task while chatting, while a phone call takes your entire attention.
Fact 2: The most expensive cost factor in the contact center is labor.
Many agents are employed to do mundane, repetitive tasks, such as asking about the nature of the call (pre-qualification), or who the customer is (authentication). Oftentimes, they are also tasked with answering routine inquiries such as “what is my balance.”
Human performance deteriorates when confronted with boring repetition, yet thrives with engaging challenges. Humans shouldn’t spend their time doing mundane and non-creative work for long periods of time; in the same way, bots shouldn’t be doing complex work, such as solving a complicated service issue, or providing a human touch to calm an angry customer. In the world of customer care and the contact center, the human and the digital employee (aka bot) can co-exist peacefully and even enhance the performance of each other.
Rather than port games or frequently needed utilities over to the world of conversational UIs (CUIs), why don’t we leave those with mobile apps where they belong, and tackle an area in desperate need of improvement?
A Simple Business Case for Customer Service Bots
Let’s do some (simplified) math on a typical customer service inquiry: “where is my order” (WISMO). On average, a call with an agent in the call center costs the business about $2.50. Now compare that to a “digital employee” having a messaging-based conversation with you; the following example goes even beyond the WISMO inquiry and shows change of delivery address:
Let’s assume that the cost per message (both directions) comes to 2 cents. With an average of 4 messages back and forth we would be at $0.08. That would mean $0.08 to do the same task with a digital employee, vs. $2.50 to complete it with a regular employee: savings of over 95%.
Finally, let’s assume you handle 5000 inquiries a day, and you can, through marketing or announcing the new option on the phone, convince 20% of your customers to try messaging vs. calling in the future. That would mean $2500/day (or $912k a year) when using live agents. Compared to $80/day (or ~$29k a year) with a bot.
Bottom line: the initial cost of developing the bot aside, following this simple model, we can save that company over $880k a year with a chatbot!
To those familiar with automation in the contact center, these numbers shouldn’t come as a surprise. IVR is known to cost only 1/10th of an agent, and ITR can easily reduce this cost further, so that even more savings can be accomplished when taking the step from voice to messaging. For most retailers, WISMO inquiries are still the number one reason why customers call them; and because of the challenge with entering alphanumeric order codes over IVR, most of these are still handled by live agents. With chatbots, what you type is what you get, so new use cases become possible that aren’t feasible or affordable with IVR.
Finally, introducing convenience through messaging can positively impact your CSAT or NPS (Net Promoter Score). Just a few points up mean a lot to many organizations these days, as customer experience is increasingly becoming a distinguishing factor.
Let’s start building bots that deliver on this vision — and then, let’s talk again about how useful a chatbot is…
Latest posts by Tobias Goebel (see all)
- How patient do we need to be with Machine Learning? - January 20, 2017
- 2016 – The Year of the Chatbot - December 27, 2016
- The Business Case for Chatbots - December 16, 2016