Editor's note: The robot is an advanced form of artificial intelligence communication that attracts and interacts with users. But what is the potential for robotics and artificial intelligence in the sales and customer service industries? What is the bottom line of their capabilities? Tobias Goebel, director of the Emerging Technologies Department of Costome Software Technologies, expressed his opinions.
Tobias Goebel: Recently, I had the opportunity to share my thoughts and opinions on artificial intelligence and machine learning. In particular, I talked with Ron Branchman of Cornell Tech and Akhi Aryan of Haptik, a personal assistant service application, about in-depth discussions about chatbots.
Since Facebook launched its smart chatbot platform at its F8 conference in April of this year, this technology has gradually become mainstream in the industry and has aroused people's enthusiasm - but there seems to be no sign that chatbots will rapidly spread. Next, share the interviews with Ron Branchman and Akhil Aryan to me.
What is the best use case of machine learning? Where are our most advanced technologies? Voice, image, or text? Or something else?
To answer this question, I need to explain it first. What I care about and are good at is the customer service area. This may be a very delicate industry. Especially when there is something wrong with a company, there will be a large number of consumers - such as the recent purchase of damaged goods, billing or invoice errors, and so on.
Naturally, these use cases are highly emotional. As a business, you need to respond to a variety of potential problems encountered by customers. Therefore, relying on chat bots to solve certain problems that require manual support does not seem to be a good idea.
Moreover, if you do not monitor, or only a small part of the supervision, only let the robot rely on the interactive content they see - for example, in the user forum - "self-learning", it will easily appear Problems, such as Microsoft’s launch of the Tay chatbot on Twitter, quickly “learned†racial and gender discrimination.
On the other hand, with proper supervision and enough training data to support it, as an artificial intelligence technology, machine learning can help to build a highly responsive and accurate information-based chat robot. For example, some robots can Supports the extraction of data from large text content (such as operating manuals).
What I would like to say is that machine learning, as a technology, has been proved to be the best industry field that can be applied, namely image processing. Google, Facebook, and startups like Moodstocks (which was just acquired by Google) all benefited from this technology, and they also showed that machine learning has a very impressive performance. However, we have now also discovered that machines seem to be beginning to take over some cognitive tasks that we believe can only be controlled by humans.
What is the bottom line of intelligent robot's ability?
From my point of view, one of the main applications of robots is in the field of customer service. It should be known that the overwhelming majority of companies are still unfortunately still using “ancient†methods to manage customer interactions. In fact, we can fully imagine what the first reaction of many customers is when they encounter problems. Yes, they will pick up the phone and call the company's customer service phone.
Businesses do not provide their customers with convenient "digital-first" interactions. In most cases, customers have to download a mobile app, open a web browser, or use voice, which seems to have become the current smartphone must Prepare the "Stay" channel, otherwise the customer simply cannot retrieve information or perform interactive operations.
This is really unfortunate, because when you need to pay a bill, when you need to check the status of an order, or when you look up an account transaction, it is almost impossible to send a simple message to a chatbot. Facebook Messenger has 900 million users, Whatsapp has 1 billion users, and basic SMS services have tens of millions of users. Today, for businesses, they have a brand-new “medium†that can interact with customers.
In instant messaging, a simple query question can be presented in the form of a simple message, such as "Where's my order?"
You can compare the chatbot with traditional service options. The latter will either "force" the user to answer some of the guiding questions on the site, either by selecting the service menu on the mobile app or by using the self-service voice on the phone. The system provides services. Now you can imagine a comparison scenario. One is designed for simple problems. The consumer-adapted, digital, automated interactive artificial intelligence chat robot. The other is the artificial seating interaction service provided by telephone. Which one is better Which cost is higher for customer service? When chatbots have the ability to handle some of the most frequently asked questions, they can reduce the pressure on artificial seats, and they can therefore be re-arranged by leaders to handle more complex and better-time customer service interactions. Some simple or even moderately difficult problems can be solved more quickly, customer satisfaction levels will be greatly improved, and then longer-term customer loyalty will be achieved. Robots can “block†some of the voice calls, which further reduces the cost of purchasing voice-enabled software licenses and the cost of telephone equipment.
How do you see the future evolution of intelligent robots? What are the major events that may occur in the next six months, one year, or two years? What is your biggest concern for this artificial intelligence robot trend?
Many companies have started and launched intelligent robot services in order to acquire customers or provide customer service. In the future, we will see that some companies have failed. In fact, some companies have already failed. Robots are not easy to develop: You need to have a lot of experience in human-machine interface design. Here I quote Amara's law: We overestimate the short-term effects of technology and underestimate its long-term effects.
It can be said that intelligent robots will be a great new platform. It will make things easier and help us all. But if robots try to do too much, or if we set unreasonable expectations for it, then consumer confidence is bound to be affected and it may no longer accept robots. There are too many hype for smart robots. We may need to calm down. Focus on developing good robots that can provide value to humans - then share our experience and tell the world where the true value of the robot is.
VIA VB