“Why a Chatbot?”
It’s a simple, three-word question I’ve posed to various entrepreneurs recently, particularly in sectors like FinTech and InsurTech. Some of these entrepreneurs have built a chatbot as a feature, others have developed it as core to their product, and a few even offer to license their chatbot to third parties (Chatbot-as-a-Service).
As a natural introvert who can find it exhausting to talk with actual human beings, I see the intuitive appeal. Yet I find that most chatbots lack a reason for their existence. Many chatbots seem to be solutions in search of problems, hammers in search of a nail. Against this backdrop I offer the following Grand Theory of Chatbots:
Bots that pull customer engagement from other channels (phone reps, websites, etc.) are marginal value at best and often a distraction. On the other hand, chatbots have tremendous value when a consumer can push an entire back-and-forth to a bot that acts on a consumer’s behalf. More succinctly: Pullbots are a dead end, and Pushbots are the future.
Pullbots are the most common form of chatbot today, encouraged by Facebook, Amazon and others. These are the bots that sit between a business and their customers. They’re available 24/7. They can answer a customer’s question and sometimes initiate action from that business. There are high profile examples and examples and examples and examples.
There is a problem here. Pullbots don’t enable new functionality for a consumer; they only provide access to existing functionality through yet another user interface. I have yet to hear of a single example where this kind of bot empowers the consumer to do something unavailable through a well-designed website or app. The first wave of deploying new technology usually skews to helping companies, not consumers, and that’s the case with the first wave of chatbots.
Pullets are incremental at best; they are never transformative. Businesses should instead focus on making sure the underlying product is simple and transparent. Businesses should ensure other channels, like an app or website, are so well executed that they moot the need for a back-and-forth discussion between customer and company.
On the other hand, consider a chatbot that works on your behalf when a delivery from Amazon is late. Or a chatbot that corresponds with a municipality to dispute a parking ticket. Or a chatbot that communicates with a health insurer to ensure you receive reimbursement for an out of pocket expense. In each of these examples, a consumer is empowered to “push” an entire back-and-forth negotiation for the bot to execute.
The most compelling opportunities for pushbots are where people have rights, but they don’t have the skills, expertise, or time necessary to take advantage of their rights. The typical consumer might not know how to dispute a parking ticket or navigate their insurance policy, but a well-coded chatbot certainly can.
Pushbots are transformative, not incremental. Rather than acting on behalf of a business, they act as an agent of the consumer. To enable the transformative impact, consumers need to invest a few moments granting permissions, access to relevant data, and ultimately authorizing the bot. But the payoff is big: Pushbots will take on tasks that consumers don’t know how to do or don’t want to do. Eventually pushbots will become not just agents, but advocates for their users.
This has big implications for any business that relies on breakage. Breakage models rely on consumers not taking action because of some hassle, confusion or other friction, which unfortunately seems all too common in sectors like banking and insurance. Technology is at its best when it reduces friction. When a chatbot removes friction and empowers users to exercise their rights or receive a benefit, that’s amazing. That’s a Pushbot!