The 3 Myths of Chatbots and the Rise of Conversational AI
Trends like AI, cloud computing, and mobile computing have driven rapid growth of chatbots. Analysis from ResearchAndMarkets.com predicts that the market size for chatbots will jump from $2.6 billion in 2019 to $9.4 billion by 2024, which translates into a compound annual growth rate of 29.7 percent.
Yet, these figures may even underestimate the chatbot momentum. The impact of the COVID-19 pandemic appears to have accelerated the adoption of chatbots for such key purposes as handling the spikes in customer inquiries.
After all, chatbots provide many key advantages. They operate on a 24/7 basis and can provide major efficiencies. Chatbots have also proven versatile for myriad use cases across a wide span of industries. Additionally, customers expect to use various digital touchpoints for interacting with companies, whether via texting or voice.
But the market remains fragmented and confusing. There are really no clear-cut definitions for chatbots, and there are many flavors for consumer and corporate environments. In fact, consumer offerings, such as those from Facebook, are looking to expand into enterprise markets.
All of these factors have generated quite a few misconceptions, which can certainly result in wasted corporate investments. Therefore, let’s examine some myths, but also follow up with encouraging trends, such as the emergence of Conversational AI:
Myth #1—Chatbots are a new innovation: Not at all. Did you know that the first chatbot was built in 1965? It was called Eliza and has since become a legend in AI history (the name was based on George Bernard Shaw’s play Pygmalion). The creator of this chatbot was MIT professor Joseph Weizenbaum.
For the most part, Weizenbaum enabled people to ask Eliza questions—via teletype—for purposes of psychological counseling. Of course, Eliza caused a stir, as some people actually thought that it was a real person! But interestingly enough, the underlying principle was to essentially repeat a variation of the question from the user (which is still a major element of today’s chatbots. You can find online examples of Eliza, such as at http://psych.fullerton.edu/mbirnbaum/psych101/Eliza.htm).
Myth #2—Chatbots are AI: Not necessarily. Granted, some clearly are highly sophisticated AI systems, such as Siri and Alexa. But these are the exception. It’s more common for chatbots to use traditional rules-based systems, relying on structures like if-then-else statements. Note that there are many chatbot creation systems that rely on this model.
A true AI application, on the other hand, would use sophisticated algorithms and process enormous amounts of data to allow for ongoing learning and improvement.
Myth #3—People don’t want to talk to chatbots: This is certainly not the case. And it’s not just younger people who are comfortable with chatbots. The reason is that chatbots can be more convenient and faster to work with. Advances in Natural Language Processing (NLP) have also been key.
It is important, however, to include a disclaimer that the system is automated in order to set accurate expectations.
Next, real value must be delivered. For example, in a survey of 1,000 respondents, about 35 percent said that the No. 1 reason to use a chatbot was to save time. Roughly 54 percent indicated that they would always use one if the savings were ten minutes.
Taking Chatbots to the Next Level: Conversational AI
Even sophisticated AI-based chatbots like Siri and Alexa are not conversational. They are command-based systems. But for AI to become optimally useful and powerful, there will need to be much more interaction.
This is what Conversational AI is all about. It’s more of what you would expect when dealing with a human. This means that the user experience will be more engaging and helpful.
Conversational AI is still in the early phases. But more cases are emerging that show considerable progress. One is Lemonade, an insurance company that focuses on auto, home and renters policies, which is built completely on AI. It has several bots that handle customer inquiries, provide quotes, and process claims.
In fact, the bot that deals with claims, called “AI Jim,” is particularly impressive. This is an area where customer care is extremely important because a person has perhaps suffered a major harm. Because of this, AI Jim was built to navigate through the emotions but also to effectively process the claim. Consider that it resolves matters about a third of the cases. Those that it cannot complete are routed to the right person, calling for much less heavy lifting.
As a testament to the power of Conversational AI, Lemonade has recently pulled off one of this year’s most successful IPOs. Shares are up nearly 200 percent with market capitalization of $4.6 billion.
Such success does not mean that any company can create AI-based chatbots on their own. This is not realistic. But the good news is that there are various startups—like Aisera—that are providing prebuilt Conversational Ai systems. And yes, they can mean moving the needle substantially in the right direction for nearly any company.