In US, consumers receive an average of 23 spam calls per month, up from 18.8 in 2017. In addition, more than half of these calls are robocalls. Now the question is, did you realize that, up till now, you’ve spent a lot of time talking with robot rather than real person?
Some may refuse to believe it, since the tone and speed of a human speaking are subtle. Guess what, they can still be FAKED, under the various AI technologies emerging recently. Check out how Google’s AI assistant can make a phone call to real person:
The technology behind
How does that work? Below is an overview of the architecture of a chatbot. This architecture is similar to a talking robot that can actually speaks.
When you speak over your phone, the system processes your voice and translate them into text. Based on the text generated, it searches for the key information in the user’s text to discover related phrases that what users want to convey, by going through the knowledge base (it also automatically corrects your grammar!). After that, the system finds the appropriate response from its’ data storage and talks back to you using a simulated human voice.
Patterns and Robotics: A Complex Reality - Data Driven Investor
Hayek's famous work 'The Theory of Complex Phenomena' (Hayek, 1964) delves into the topic of complexity and asserts…
Who will be impacted
It will definitely impact the service industry.
Because, making a phone call is essentially a process of one sending a request and the other handling it, which is exactly what service is. Accordingly, the technology will benefit both sides significantly.
From the company’s perspective, the AI chatbot will greatly improve productivity, save money, and reduce work. For example, nowadays the customer service receives thousands of similar requests each day, and those repetitive calls keep on challenging the patients of our “human receiver”, making them frustrated and even pissed off. However, this can be avoided by using the AI chatbot, as machines have infinite patience and are best at doing duplicate work. Moreover, it reduces the cost of hiring addition people to endure this process.
From the client side, we can send out request efficiently and effectively. Back to the Google Assistant mentioned above, the assistant is capable of doing various tasks ordered by us, such as booking a room, asking for weather, etc. Then we wouldn’t have to spend time on communicating with the service side or typing our question into the search engine. Therefore, it would bring conveniences to the client.
How can we adapt
As for client, adapting to such technology is quite easy for us, as it provides us with pure benefits and no harm.
As for company, when the automatic process has been applied to the workflow, the sad truth is that it may cause unemployment as there are still people who relies on phone calling for living.
Here, I see two potential solution to address this problem.
First, in order to automate the process, we need people to construct answers for the training data (voice of client), so that our model (AI chatbot) can learn from it. To be more specific, the chatbot itself cannot understand anything from pure human’s voice unless someone tell it what it means and how it should reply. And this “teaching” process requires huge amount of manpower, and we need lots of people to perform task like this.
Second, as there are still spaces for improvements in AI chatbot, so the customer service cannot fully rely on it. As a result, we still need people to monitor the conversation between human and robot, and to intervene into the talk when necessary. Take autonomous driving as an example, there is always a backup security driver sitting on the driver’s seat during road test, in order to prevent malfunction of the automation system. Similar to that, we should always have someone to ensure that the request of the client is being properly handled.