CHATBOTS ARE NEARING THE END OF THE PHOENIX CYCLE
It seems that with the release of any new technology, the same pattern emerges: The Phoenix cycle. First, a technology is created, usually by a few earlier adopters who predict a trend and have a specific market need. Second, the hype train. Everyone jumps onboard for the technology, talking it up as the next big thing and embracing it wholeheartedly. Take the Long Island Iced Tea Corp and blockchain hype. This iced tea company literally changed their name to the Long Blockchain Corp. What does an iced tea company know about the blockchain and why is it in their name? No clue. The new technology gets applied to everything without much thought as to why. Third, people start to realize that a specific technology is a square peg trying to be stuffed into a round hole. Once people realize that some new technology isn’t the answer for all technical needs, the doom and gloom prophecies begin. Just check out TechRepublic’s “Chatbots are dead. A lack of AI killed them,” or WIRED’s “Facebook’s Virtual Assistant M Is Dead. So Are Chatbots.” This doom cycle seems to last about half as long as the hype cycle but something funny always happens towards the end of the cycle: the phoenix companies emerge from the ashes. These are companies where the technology has true market fit. Whether by luck or by good planning, they have real business cases and value propositions. We are now entering the Phoenix stage for chatbots.
When chatbots first launched, people saw tremendous opportunities with them and tried to apply them to many different industries. I can’t even remember how many news chatbots were released. This is a perfect example of misusing great technology. Using a chatbot to read the news just doubled the amount of time it took to consume information. Flipboard is the ideal interface for browsing the news quickly online. If you’re searching for a specific topic, nothing beats the speed of Google News. People might want to chat with others about the news, but they don’t want to consume their news in a chat interface. It just doesn’t work well. Yes, getting a breaking news alert in a chat window is cool, but how is this different than any other type of notification? Chatbots by their very definition are about two-way communication and not just broadcasting messages. People have been accustomed to reading, watching or listening to the news. To make chatbots work here, you would have had to wholesale change the population’s behavior. Not a good fit.
Another bad use of chatbots is delivering weather information. Chatbots like Poncho were cute (they closed down) but don’t deliver weather information in a better manner than existing technology. If I’m on my Mac, I can click on my notification center, and within seconds I’ll get a beautiful graph with weather information for the next five days. If I’m on my phone, I usually click the widget on my home or lock screen that takes me to Weather Underground. It takes seconds to reach, and the information it presents is more visual with more in-depth content than what Poncho provided. If I’m in my home, I’ll just quickly ask the Alexa voice assistant the weather and get the answer I need. Maybe they will have success with their Slack integration or something else, but I don’t see chatbots being a mainstream long-term solution. Weather isn’t something that requires two-way communication. There is no gain in efficiency or ease here. Chatbots are not suitable for routine consumption of information, especially non-targeted information, like browsing.
So, where then do chatbots fit in? I believe chatbots are here to stay for the long-term because there are a lot of complicated topics that require a conversation. If you need to chat with someone about a specific issue, a conversational interface might make sense. It’s not a single sentence update but for guiding people through a complex environment. These complex environments are where you can derive greater efficiency and a better member experience through chatbots. If you search through some of the most difficult and dense subjects like banking, employee benefits, or education, you will find the best examples of chatbots. People historically have relied on conversations with others to navigate these subjects, so you don’t have to fundamentally change behavior to get people to use chatbots. You make that behavior easier and faster. Trying to help people navigate something like their health insurance has always been very complicated, that’s why we put a chatbot at the center of our healthcare guidance platform. There are many different variables to consider on top of a confusing system. You combine this complexity with the ingrained behavior of having a conversation, and you have a field well-suited for chatbots.
Chatbot technology will only continue to improve over time and be able to provide a deeper, more personalized level of service. It’s still early days but chatbots like language learning bots from Duolingo and free legal services from DoNotPay show the incredible value of chatbots.