Artificial Intelligence, Chatbots and Real Business Values
Our chatbot team based in Hong Kong and Taiwan, has been actively doing market education in Asia since April 2016. The terms “artificial intelligence” and “chatbot” are now hot to consumers and marketers. However, unlike a lot of software firms who don’t have experience of launching real chatbots, we gained insights from our successful and failed launches.
Don’t forget the original purpose of making chatbots
Everyone is using WeChat to do everything in mainland China, much more common than standalone websites and mobile apps, because of the all-in-one ecosystem and convenient P2P payment gateway. Such big success of penetration is something we wish to replicate to the whole world, through different messaging platforms in different countries. That’s why Mark Zuckerberg introduced “chatbot” as the new UI for B2C communications on Messenger, one of the most popular apps that people are using everyday, enabling native utilities without need of downloading new apps.
“Be as digital as your customers” is the message we have been delivering to clients with Facebook as seen from HSBC Digital Summit, because we believe making chatbot is the new digital strategy to engage young customers in the same way as building websites in year 2000.
Messenger bot vs chatbot
Mark’s speech impressed a lot of people, especially the brainstormers from innovation labs of big enterprises, because the prefix “chat-” sounds game-changing to CEOs. And then we see some of them invest millions of dollars without launching a working prototype after a year, possibly due to disappointing chatty experience other than discussing about roles and responsibilities.
This year, the term “chatbot” was rephrased as “Messenger bot” on stage of F8, in order to remind you Facebook’s original purpose was to enable robotic automation using conversational UI on everyone’s Messenger. The success of WeChat in China is not about chatbot, so we don’t expect Messenger bots will be widely adopted because of being chatty.
Artificial intelligence is not only about natural language understanding
I see a lot of chatbot makers keep emphasizing their outstanding performance of natural language understanding (NLU or NLP), demonstrating how artificial intelligence is applicable to different customer service scenarios. First of all, we believe the smartest PhDs are hired by Google, Facebook, IBM, Microsoft and Amazon. Inventing a smarter NLU than Google is as hard as moonshot, so most vendors are utilizing the same budget-friendly NLU solutions from these five big players. The differentiation of NLU capability among vendors therefore is not that large. The choice really depends on execution and business value understanding.
Next, I’m not going to elaborate the better way to implement NLU although our technical team at Whatsquare spend lots of effort everyday to deliver the best natural user experience. What we should really focus on is to bring immediate values by adopting artificial intelligence. NLU is an important element of AI, but the ROI of answering customers’ questions at less than 30% accuracy is nearly zero, especially when the enterprises have existing human CS team able to answer accurately to long questions having multiple intents, when small merchants can efficiently handle less than 10 inquiries per day.
Who wants to spend a lot of money to lower satisfaction rate from customers?
What are your real business values?
The biggest values of artificial intelligence, nowadays, I can find is about smart automation, about how we can mimic the best salespersons’ tactics, and the most professional consultants’ advice, as well as suggesting the next best actions by looking at data patterns. These have never been scalable before without AI but very effective to convince target audience. Therefore we set the objectives of our chatbots in line with the ultimate goal for business, which is to acquire new customer and retain loyal members.
In real life, AI can be used for personalized product recommendation or behavioral user segmentation, which are very suitable to be implemented in active conversational way, comparing with passive websites and mobile apps. For example, our global beauty brand clients find chatbot more valuable than other digital platforms as a personalized and sticky communication channel in every part of consumer decision journey, from free treatment booking to product replenishment.
Yes, we design every chatbot starting from analyzing consumer journey. One business, one bot, one journey, comes with unlimited new possibilities. Hope this methodology would be insightful to chatbot makers who are facing difficulty to achieve high expectation.
