CHATBOT: THE INSURANCE AGENT’S FRIEND (AND PART-TIME RIVAL) – EPISODE 047
In this episode of The Digital Broker, Steve and Ryan talk to tech entrepreneur Duke Williams about chatbot technology and how it applies to insurance operations. By listening to this episode, you will learn:
- How chatbots basically work and why they’re useful inside an agency
- Why some chatbots are popular and others aren’t
- Where a chatbot does and doesn’t threaten an agent’s job
Duke has found that many agents and brokers misunderstand or underestimate what a chatbot is and what it can do. It isn’t simply a widget that resembles a conversation between computer and user. Fundamentally, a chatbot is an A.I., rules-based conversation within a specific context. (5:57)
Despite the strides in A.I. technology, most chatbots are pretty simple creatures—Duke likens them to four-year-old children, but even that might be overstating it. Four-year-old children can think independently and generate new ideas. A chatbot is limited by what a human tells it to think and how.
We call these limits a chatbot’s context. (7:11) Incidentally, more context is not always better. Siri is great at turning on the lights and everything, but try asking her to help you fill out a CERT. Nobody’s bothered to optimize Siri in the context of insurance, so you’re not gonna get very far. For practical purposes inside your agency, the best chatbots tend to be the ones that are most specialized, but this presents its own design challenges. Duke recounts some of the ones he encountered at 10:16 of this episode.
If a chatbot can simulate a conversation between a client and an agent, won’t it eventually replace the agent? Yes and no. What a chatbot is most likely to take over are basic tasks, like filling out forms. Actually, a chatbot is an agent’s friend here. Clients do a better job filling out forms that have been optimized as fluid and responsive conversations over static documents. Moreover, there’s more to being an agent than filling out forms. Agents advise the client through the complexities of life’s risks in a way that chatbots still can’t. (13:08)
Far from gunning for the agent’s job, chatbots can do wonders if you give them a chance. As with many new technologies, chatbots are supposed to save you time and make your life easier by inheriting the tasks you’d rather not do or aren’t good at. There are coverages, or parts thereof, that are too arcane to spend a great deal of time on but straightforward enough to delegate to a chatbot, allowing you to focus on the coverage you’re awesome at.
But who says a chatbot has to be client-facing all the time? A chatbot is supposed to guide anybody through a complex process, and sometimes, the people who need the most help with that are the agents themselves. Are there processes within your day-to-day that require an agent to put everything on hold in order to source every step of the process from various sources? What if you had a chatbot that had been designed to guide the agent through every step without having to leave the desk?
Yes, you can design your own chatbot to fit your agency’s needs. Chatbots make amazing repositories of an agency’s collective knowledge, but only if somebody feeds them the information, creating an incredible opportunity to engage your employees, especially your senior ones. When an agent retires and leaves the agency, so does the agent’s intellectual capital. A chatbot can keep and preserve that knowledge, redistributing it to all who need it for years to come. Here, your senior employees aren’t simply qualified to feed the chatbot; they are probably the best at it, with their years of accumulated wisdom. Encourage them to get involved and keep up with the chatbot on a routine basis—chatbots require maintenance and periodic fine-tuning to provide the best responses to all the questions that they’re asked and all the ways they can be asked. (20:33)
We’re seeing these agency-facing benefits of chatbots being leveraged to great effect. Thanks in part to Duke’s efforts, chatbots are integrating into dense training material, where jumping from section to section is tough. By overtaking the help/FAQ feature, a chatbot allows the trainee to ask questions about any part of the material at any time. An added benefit of chatbots is that everything is a matter of record, which comes in handy during dispute resolution. Erring, after all, is human. Agents sometimes give one type of advice and a different type the next day. Clients forget or misinterpret what they heard and when. A chatbot’s output, on the other hand, is consistent on a day-to-day, client-to-client basis and contains precise information about everything that was said and when.
All this talk of chatbots makes the hosts curious about conversational computing. (30:09) It’s too soon, and a little nauseating, to imagine a roomful of account managers talking into their computers at once, but conversational computing is gaining ground in the office environment. Chatbots are maturing into manager-bots by acquiring insights into productivity baselines and making sure employees are hitting benchmarks and targets. The turning point, Duke says, is that computers are beginning to talk our language rather than the other way around.
You’re probably talking to enough computers, be they chatbots, manager-bots, or whatever. How about taking the time to talk to people who are passionate about insurance like you are? You can meet a lot of them in our Digital Broker LinkedIn group, where we take questions about anything you hear on and off this podcast. What’s your experience with chatbots? Does the prospect of a manager-bot thrill or worry you? Come tell us all about it here.