Wow! What a blessing over this holiday season! I just had the best customer service exchanges online ever – five stars! Guess what? They weren’t human.
You’re frustrated. You promised to finish the design you’ve been working on by the end of the day. But, it’s 3:15pm and you have just spent what seems like the last hour trying to get the software to accept your last inputs and give you a preview before you make it final. As you curse under your breath, your buddy leans over and suggests: “Why don’t you just punch up AVA?” Ugh! The idea of reaching out to the vendor’s customer service department – this is not your favorite activity. That‘s why you have been going through the user manual and trying to solve the issue yourself.
But, the clock is ticking so you punch up the software vendor Autodesk’s customer service online. Boom! In a window on your screen appears – a person. (That’s her above.) AVA says hello and asks how she can help you. You quickly describe the problem and she smiles and assures you she is pretty sure she can help. She asks you to also turn on your camera so she can see you, too. Ordinarily you are not a big fan of using the camera but this seems so much like a Skype conversation with a person you flip it on.
AVA – looking surprisingly sympathetic – then launches into some questions and elementary suggestions. Your response is: I tried that – over and over again. Your frown deepens and your frustration increasingly creeps into your voice. AVA sees and hears this and stops. Looking apologetic she says: “I am sorry, but I can see this isn’t helping. Have you already gone through the trouble shooting guide in the manual?”
Caught off guard, you tell her yes and confide that you are under some time pressure. She sympathizes and suggests a new approach. She jumps ahead to recommending some keystroke commands, which also appear on the bottom of the screen. You’ve seen those in the manual but were not sure of their effect. AVA nods empathetically but encourages you to go ahead and try them. You do and voila! The last design elements are accepted and the program asks if you want to go to “preview”. You get a big grin and AVA smiles in response, seeing you’re happy.
Welcome to customer service as we move into the end of the second decade of the 21st: efficient, effective, empathetic and not human.
Who among us have not interacted with a contact center? How did you like it? If you were lucky it might have been just OK. 80% of enterprise executives believe that they offer superior customer service. Sadly, only 8% of their customers think so.
The call center industry is huge and targeted to reach worldwide over $407 Billion annually by 2022, but this is a tough function to run. It is very people intensive with human capital being its biggest expense and key determinant of customer satisfaction. World wide contact centers employ over 3.5million people. One in twenty-five people employed in the US works in a call center. Staff attrition is awful: 15% bail within the two week training period, 32% are gone within the first six months.
But a big shift is underway – a move away from humans staffing contact centers. Gartner predicts that by 2020 two-thirds of customer interactions with the enterprise won’t involve people. The contact center model has been morphing over the years. Initially, it was primarily done through voice (i.e. a telephone call) and in an attempt to automate these interactions companies introduced IVR’s (Interactive Voice Response) but those systems were limited in capabilities. How many of us have dealt with these only to be frustrated and scream “operator”, “customer service”, “representative” or worse trying to get a person on the call?
As the generations have shifted, customers have been comfortable to move beyond voice. Today chat has expanded to almost 63% of the interactions. Other kinds of messaging like text; email and social media have also been added. Being text based (a machine-readable format) managers sought computer assistance to both reduce costs and improve performance. The result was the rise of the chatbots.
Chatbots are algorithms that have been programmed to respond to customer inquiries based upon large databases of what customers have asked in the past and what responses met their needs. The initial ones were primitive but now they are increasingly sophisticated in their apparently personal response.
Here is a good example of these from the world of education. Georgia Tech may have hit on the way to address this need for personalization. GaTech has a very well known and increasingly popular online master’s in computer science. It’s so popular that Ashok Goel, the lead professor in a course on AI, couldn’t deliver the personal experience (read: customer experience) he knows students deserve because he couldn’t leverage his Teaching Assistants far enough.
His insight: most teachers know that most students ask the same questions year after year (same lesson for most inquiries – the 80/20 rule reigns) as they wrestle with the material. Doctor Goel thought why not build an AI to help. The result was Jill Watson (yep, the “Watson” of Jeopardy fame). “She” was so successful that most students did not know “she” was not human.
In a virtuous loop, better speech recognition was also driven by AI making IVR’s much savvier and improved the customer experience. The likes of Alexa, and Siri coupled with that 80/20 rule now makes the need for people attendants reduced to just the unusual inquiries that the AI’s cannot handle.
Amazon CTO Werner Vogels points out the future of computing will be driven by human-centric interactions—inputs that are more similar to how we communicate and interact with one another in everyday life.
Research has shown that only 7% of any message is conveyed through words. We actually convey and read so much more information from facial clues, tones and bodily language. How to deal with this is the new challenge. Welcome to the AVA (Autodesk Virtual Agent) at the beginning of this post. The 3D design software maker Autodesk is building this online help service that allows people to interact with what sure looks like an actual human.
Autodesk started with an AI chatbot, IBM’s Watson’s Conversation API then added the exceedingly lifelike face, voice, and set of “emotions” provided by a New Zealand startup called Soul Machines. Finally, the company is experimenting with the Watson Tone Analyzer to read the customer’s mood. (For a deep dive check out this Fast Company article.)
Imagine where this will lead: avatars customized to emphasize a brand like “Flo” for Progressive Insurance. Maybe getting your taxes done online could be a whole new experience. We’ve already highlighted one example in education but why stop at just graduate school? Why not K-12? What possibilities can you envision in your world?
Brave New World? – No, just the usual tech replacing labor for simple, repetitive tasks – after all we don’t use operators to connect telephone calls anymore.
Image Courtesy of Soul Machines