Can human and artificial intelligence be combined?

Today, more than ever, artificial intelligence is at the heart of all conversations revolving around online customer experience. A few years ago, chatbots and artificial intelligence were perceived as not efficient enough. But, since the F8 conference when Facebook announced the launch of their own bot store on the Messenger platform, they are seen as a powerful customer experience tool. Chatbots and AI evolved over time and now, combined with human assistance, they are the future of online customer service.

Read more: Chatbots and artificial intelligence: where do we stand?

Artificial intelligence and its limits

Today, even the most advanced artificial intelligence has an IQ score roughly equivalent to the one of a 4-year-old. And of course, companies are not likely to entrust their customer service to a 4-year-old.

Deep knowledge and emotions

But the most important problem of artificial intelligence is that, as it’s indicated in its name, it’s artificial. Humans are driven by their emotions. And emotions are precisely the most complex issue for AI as they are really difficult to simulate. Consequently, the most important challenge for artificial intelligence is to acquire deep knowledge. AI has yet to enrich its knowledge. For now, deep knowledge is owned by experts who exchange experiences and best practices with their peers. That’s what makes their knowledge so unique, it’s fieldwork!

Artificial intelligence, a powerful comeback

Maybe you’re wondering why chatbots and artificial intelligence are making a comeback if they are so flawed. There have never been so many companies focused on chatbots such as Polly, Sensay, Dexter, Interactions, Inbenta or AdmitHub. And all these companies focus on artificial intelligence and they’ve all raised millions of dollars last year. Artificial intelligence has benefited from the launch of Facebook’s own bot store on their Messenger platform. If Facebook is betting on this trend, it’s at least worth considering. And some of the chatbots on the Messenger platform turn out to be quite efficient!

CNN’s chatbot on Messenger

CNN has implemented a bot on the Messenger Platform. It’s one of the most commonly used examples of chatbots. When you’re logged in to Messenger, through this chatbot you can have a direct access to breaking news with menus in the form of lists and hypertext links. This way, you have access to a summary of the content.

What changed since the beginning of AI?

I personally think this type of experiences we’ve seen with the CNN bot can and must be improved. Did you know what stopped the development of the internet in France? The minitel. The user experience on this device was simple: you had the capacity to visualise different options in the form of lists, you had access to different sections in order to display content in the form of a text. What changed since then? The touchscreen interface, the loading speed, the resolution, but not the different options suggested to users. 

Is fully automated AI possible? 

We are far, far away from the potential announced for chatbots. Companies are doing experiments about fully automated AI but they are not all conclusive. AI is not ready yet to be fully automated and this resulted in a bad experience for Microsoft. Tay, Microsoft’s chatbot turned into a big controversy for Microsoft in 2016. The chatbot was a network of virtual neurons which learned from human interactions just as a child would do. And just like children, if you used foul language, Tay learned from it and repeated what you said. Within 24 hours, Tay became a fascist and revisionist AI. It was then disconnected by Microsoft. 

Artificial and human intelligence: bound to be opposed or the perfect match?

The exoskeleton has the answer to this question: it offers way superior capabilities than the machine on its own. Machines and humans are a perfect match because they are complementary. Let’s take an example in online customer service: 20% of conversations can be completely automated, 50% are supported by humans in charge of validating or correcting answers offered by bots and 30% of them are handled by humans alone. Artificial intelligence can take care of “low value” questions, very frequently asked questions and can guide customers towards the most suited service to meet their demand. This way, professional agents can focus their attention on questions where they will really make a difference, and add value. They can give more accurate and personalised answers to “high value” questions. They can intervene on a wider variety of questions and less recurring ones for which their expertise is valued. 

Read more: Are Chatbots the solution to a positive customer experience?

Artificial intelligence and human agents, the perfect match for online customer service

At iAdvize, we combine chatbots and human agents to enhance the customer experience our customers offer to online shoppers. This way, brands can automate certain aspects of the interactions they have with their customers and prospects and human agents can focus on the expertise they can offer to their customers and prospects.

Artificial Intelligence iAdvize


Julien Hervouët

CEO and founder of iAdvize

Leave a Reply

Your email address will not be published. Required fields are marked *