Chatbot & The Rise of the Automated Insurance Agent
The world’s first Chatbot was called Eliza. “She” was a computer program written in 1966 by Joseph Weizenbaum. And they’ve been used ever since to talk with humans. So why is it that in the last couple of weeks, the world seems to have woken up to Chatbots? Rick Huckstep goes in search of the answer and talks to Spixii, the little blue parrot that sits on your shoulder.
A couple of months ago I was at a London insurance market meeting. It was mostly attended by brokers and underwriters and the subject was the London TOM.
During a roundtable discussion I mentioned an article I’d just written about big data, artificial intelligence and machine learning. I said as much as 80% of insurance underwriting will be automated before long.
Sitting opposite me was a London market broker. He reached inside his jacket pocket, pulled out his pen and said;
“Nothing writes business faster than a Mont Blanc pen!”
And he was deadly serious.
Now, I’m not trying to make fun at his expense. How could he know any different, he wasn’t a subscriber to The Digital Insurer at the time (although he is now!).
In the specialist insurance market of London, this mind set may have held the market in good stead since the days of the quill pen. But the rest of the world has moved on and embraced technology.
And this is the point. That technology is rapidly catching the insurance industry up. It’s just that some in the insurance industry don’t yet realize it!
Whereas the banking focus of Fintech was all about “disruption”, the digital innovation focus of InsurTech is about “rapid evolution”. A great example of this is the Chatbot, which is short hand for an automated insurance agent in our market.
Real-life Humans talking to Real-life Computers
What is a Chatbot?
A Chatbot is a computer software program that is able to communicate with humans using artificial intelligence.
And, to the extent that humans don’t realize they’re talking to a computer program. The tech has been widely used in the insurance industry for over a decade and a great reference site is Chatbots.org.
Interestingly, most of the Chatbots are given names to make them more human-like!
Magda is the Chatbot for Link4, Poland’s first motor insurance policy sold online.
Allie is Allianz’s online assistant available 24/7.
Mia answers customer’s insurance queries for Co-op Banking Group.
Arbie “works” for RBC Insurance in Canada.
Nienke is in the Dutch market talking to NN’s customers about insurance.
Marc represents Credit Agricole’s health insurance offering in France.
Hanna has been around since 2003 for the Swedish Social Insurance Agency.
The Rise of the Automated Insurance Agent
So, why now?
A couple of weeks ago, at Facebook’s F8 conference, one of the major announcements was that they are opening up the Messenger platform to Chatbots.
Or this post about Facebook’s 10-year plan to basically be the entire internet!
The relevance to the insurance industry is simple
Opening up its Messenger platform for anyone to develop and deploy Chatbots also opens the door for the automated insurance agent.
It allows insurance firms to deploy distribution, claims and customer service straight into a platform that has 900 million regular users each month and supports 60 billion messages every day!
Spixii – the little blue parrot that sits on your shoulder
Having just completed the Startupbootcamp InsurTech 2016 cohort and previously been with the Startup Next accelerator, I turned to the founding team at Spixii to help me understand more about the impact of Chatbots in insurance.
I sat down for coffee with two of the three Amigos behind Spixii; Renaud “who loves insurance” and Alberto “who eats data”. Missing, was the third Amigo, also named Alberto, “the man who talks to machines”.
First, I had to ask them “Why the name Spixii?
“We were looking at what to call ourselves and initially we thought of ARA by combining the first letters of our name. We thought this would be a really cool name for our AI Chatbot platform.
“We found that ARA is a species of parrot and we liked the idea of branding ourselves with a gender neutral parrot that sits on your shoulder and talks to you. Which led us to Spixii, which is a small blue McCaw.
“And since blue is the color of trust, we chose Spixii.
We would like to mention here that customer feedback is at the heart of every operation of our business. In fact the decision of the logo, brand, final parrot name was all led by tests and customers interviews who made the final decision.”
Why “now” for the Chatbot?
First, people already text a lot. There are billions of messages sent every hour. Humans have substituted verbal communication (human to human) with written communication (human to machine to human). From the very beginning of computers, humans have communicated in our language – If X, then Y – which is then interpreted into machine language of Ones and Zeros. Humans are already chatting with computers and Chatbots make it possible for the machine to chat back!
Next reason, we are in App overload. Think about your own smart phone. Research shows that we only use about about 5 regularly, and half of these are social media apps. Most of us are at our limit for apps and dont need another one. Chatbots don’t need an app to engage with customers.
We already use them. The original Instant Messaging platforms used very basic Chatbots to respond to text. So the chances are that we’ve all used them sometime along our digital journey and just not know about it.
Finally, evidence suggests email has run its course. It’s great for sharing information but horrid at conveying understanding. Which is why alternatives to email, such as SLACK, allow humans to communicate in a more responsive way than email. People are more engaged with a digital chat experience than they are with an analogue email exchange.
If you, like me, are on the wait list for Denis Mortensen’s AI personal assistant AMY, then you’ve already recognised there’s a better way.
What is Spixii?
Spixii is a tech business built by insurance experts which starts by selling off the shelf products. It will be the brand that customer’s connect with as they distributes insurance products using their automated insurance agent, aka a Chatbot.
I was fortunate enough to play with a private beta tester of the Spixii platform recently. This was set up to emulate a travel insurance quote and buy process. It was a simple A/B test against a classic travel insurance form.
Which is about as simple as it gets when buying insurance. And with Spixii, the Chatbot behaved like I was in an online conversation with an real-life insurance agent. Which I was, only it was a computer program and not a real-life person.
I was talking to a machine…and enjoying it!
Eventually, Spixii will engage with customers in a dynamic conversation. This will enable greater levels of personalisation than can be achieved using web forms. Which then takes us down the path to Spixii performing automated underwriting functions based on dynamic data rather than the rows and columns limitations of today’s actuarial spreadsheets.
For example, when I beta tested Spixii I used a trip I’m about to make to the Le Mans 24 hour race in June. Because of limitations in the back-end systems, all I could “buy” was a single product, single-trip European travel insurance plan.
However, with Spixii the customer engagement could be highly personalized and interactive.
The Spixii Chatbot could ask the question; “where are you going?”.
I would reply “I’m going to Le Mans next month”.
“Do you mean the 24 hour race on the 18-19th June?”, replies Spixii Chatbot. “That sounds cool, Rick. I see the long-range weather forecast is looking good for the race. Let me ask you about your travel plans. When will you go out to France?”
“When will you return?”.
And so on, you get the picture. The point is that engagement for the customer is, quite simply, much more engaging than filling out a static web form online.
For the insurer, the risk assessment is based on better levels of information specific to the trip. Which means that my premium is going to reflect specifically what I need to be covered for. I have no gaps and the policy is less likely to be over or under-covered.
For a small premium insurance product, this level of customer insight, engagement could not be afforded without a Chatbot. And it runs 24 x 7!
To engage with Spixii, click below.
To see Spixii’s pitch at Startupbootcamp InsurTech demo day, watch this video.
The author, Rick Huckstep, is an advisor, speaker & InsurTech thought leader and editor of InsurTech Weekly for The Digital Insurer.