It’s a channel. Nothing more, nothing less.
People can exchange information by phone, fax, website, app or a chatbot. Nothing makes me crazier than when a potential customer asks, “I am being approached by so many chatbot companies, what makes you guys different?” This can mean only one thing – it is time to separate the wheat from the chaff.
Now let’s take a step back. What does Elafris do? Do we build chatbots? Or AI assistants? And how much AI is in it? To get a proper answer, we will share with you the story behind Elafris. It was late March 2016 when I got pulled over traveling southbound on I-280 and was asked for my driver’s license, registration and a proof of insurance. Driver’s license and registration were handed over in no time, but the proof of insurance–this is a whole different story. You see, me being a tech guy, I stopped printing out those silly pieces of paper many years ago, and with the GEICO app on my phone, I felt prepared for any scenario. Too bad that the last time I opened that app was…umm, never. When I launched the lizard’s app, the first thing it did, you’ve guessed it, started updating itself. This was when I started to sweat. Those were the longest three minutes of my life. The cop was cool and said, “Take your time. We are not going anywhere.” Once the app finally updated, I launched it hoping for the best. And as they rightly say, hope is not a strategy. The app asked for my username and password. My heart sunk. I am using strong, unique passwords for all of my logins. Being far away from my computer meant one thing…hack into my own system to figure it out. Five minutes later I’ve logged into GEICO app to demonstrate the digital proof of insurance. All together it took about eight minutes and at least ten pounds of sweat.
In the following days, I talked to good friends about how using messengers for business can help solve the problem I’ve experienced firsthand. We came up with the idea of building a virtual insurance agent capable of pulling a proof of insurance from a policy admin system and displaying it right in the messenger in a real time. It was a very basic implementation of a chatbot. Was there much “Artificial Intelligence”? None whatsoever. Any machine learning? Zip. Just simple automation using a more convenient (for me) communication channel, a messenger.
Let’s say there were five more features just like the one above. Would it make a chatbot an “AI solution”? Absolutely not. Again, this would be a basic automation. Furthermore, I’d call it IVR-like scripting. Nevertheless, it is still a chatbot. Would it be useful? Perhaps. Let’s take claim the submission process as an example.
Having a good script that carries the customer through the process means reduced errors via structured guidance through the unfamiliar process.
Granted, each data input block would have to be either a menu selection type or predefined. Also after each selection, there would have to be a confirmation to complete a section data entry.
Obviously, this would be a poor user experience, and instead the chatbot should be able to take the following data as an entire line input, break it down, and extract required parameters. Then, if the user wants to change one of the parameters, the chatbot would do only that change and not drag the entire conversation back to square one. The user conversation can look like this
Unlike the first scenario, the “AI” in the chatbot could figure out the vehicle details, and type of damage. It also allowed going back to change one detail without dragging customer through an entire Q&A process. Such a chatbot can run inside a web page, smartphone app or a Messenger (e.g. Facebook Messenger, Skype, or Slack). Early implementations focused on improving customer service, but more advanced ones work to market and cross-sell other insurance products.
This is what we do at Elafris; we build and train Virtual Assistants to understand insurance-related questions and to not simply answer rudimentary questions like “What is a deductible?”, but act as a true insurance broker, suggesting options that best fit the customer – “Based on the value of your vehicle and the fact that it is still financed I’d recommend adding comprehensive coverage and $500 deductible.”
To see how Elafris Virtual Insurance Agents can help your company click here
to schedule a free online demo.
In the next article, we will dive deeper into the difference between good and bad chatbots.