قالب وردپرس درنا توس
Home / Insurance / How AI changes insurance, with Lex Sokolin

How AI changes insurance, with Lex Sokolin




Futurist and fintech entrepreneur Lex Sokolin explains the difference between automation and artificial intelligence (AI), and how AI converts the insurance value chain-from chatbots to claims.

Highlights

  • Automation is the process of translating human process into machine process. It is programmed from the top, with a known workflow and known results.
  • Artificial Intelligence (AI) is the digitization of human intelligence into machine intelligence. It requires extensive data, entered into a mathematical algorithm, to create correlations between thousands of different parameters. It enables large-scale decisions, but unlike automation, the workflow and results are not known.
  • Sales and receivables agents are two examples within the insurance chain where automation and AI could be applied. But in general, chatbots have not been very effective in replicating human interactions.

How AI Transforms Insurance, with Lex Sokolin

Welcome back to Accenture Insurance Influencers podcast where we ask industry leaders about trends and technology design of the future of insurance: self-driving cars, fraud detection technology and customer centricity.

Lex Sokolin is a futuristic and fintech entrepreneur. In our latest section, he explained why banking and management governance trends can hold valuable insurer lessons, especially when it comes to working with non-insurers. In this episode, Lex shares some AI myths and looks at how the AI ​​could be applied to the existing insurance chain.

The following transcripts have been edited for length and clarity. When we interviewed Lex, he was Global Research Director at Autonomous Research ; He has since left the company.

I think many people confuse the AI ​​application of AI – with the application of automation. Can you mark the difference between the two?

If you think of digitization or automation, for me who divides two types of vectors. The first is a human process for machine process, something a person does in a manual manner, in a workflow. Take it and put it in software.

We have experience of this every day: Remember to go into Excel and enter a mathematical formula. You define a rule set according to which software will calculate something. Or you go one step further and say, "Let's build account opening software." Instead of a person entering the office and filling out paperwork, you can capture it on a mobile platform.

Taking these data and filling out forms, they are all programmed at the bottom. We know what the workflow is. It is easy enough for us to outline it and convert it to "about this, when it" regulates and then the result is completely deterministic from where we started and what type of data we added. We know how it works. We can transform the code and understand what happens very easily.

The other way of digitizing can be from human intelligence to machine intelligence. And in machine intelligence, there are different ways to create results that feel like intelligence, which feels like there is an element of judgment against it. Who is popular right now is machine learning enabled by mathematics called neural networks.

What neural networks work very well is to solve a problem in a probabilistic way to create an intuition for anything. If you are a human looking at a picture of a cat, you know that it is a cat and not a dog, and there is a process in our brain through which it happens. The image of a cat does not matter to a computer if it is not converted to data. You need millions of versions of the data cat to be aggregated and fed into a mathematical algorithm that can create correlations between thousands of different parameters to say "this is more likely to be a cat" or "it's more likely to be a dog."

AI is still software. It is still a tool, but the basic element is not the top logic of "about this then." The foundation is massive datasets where the software sits or is trained and the basic data sets came out of the Internet. And when you have these datasets you can apply these different mathematical algorithms on top. You can essentially put in a rule set how a person would make a judgment, and then you can lift that judgment and you can connect it to a software process. And then, on a large scale, you can do things like deciding whether someone should get more credit and get their next loan. And with every new information you update it.

Much of this came out in advertising. Amazon is very good at giving you suggestions on what to buy next, and Netflix and Spotify know your tastes in video and music in the same way. And for insurance purposes, there are many different ways that AI can be used in the manufacturing layer, on the operating debt layer, in the portfolio management layer, as well as in customer distribution.

So, two very different worlds. Automation is "about this, then that" command, a Soviet central planning world where you define all the deterministic results. And then, the AI ​​world is probabilistic, based on existing data you train the neural networks, and it's much more like codifying a human intuition and then using it on a scale.

Finally, one of the things that plagues the idea of ​​leadership in this space is painting with a very wide brush. Pictures of cyborgs and various network diagrams to make it feel futuristic. These things [AI] end of the day, are just a set of human tools that people developed to become more effective, to scale their thinking and simply do more. Although it sounds threatening or very ambitious, I do not think that AI is different from the clouds, electricity or wheel or fire or language – or any of these fundamental things in human development.

So in the case of AI, what solutions or types of solutions are the most mature?

I would say that the most mature parts of artificial intelligence are those built by the big tech company. The great techniques are motivated to recreate a lot of human minds. They want to find out how to provide products and services to people in a way that is intuitive and selected by the people in these platforms.

What I mean by that is the vision, the hearing loss, the ability to make speech. These are things that are very mature when it comes to the technology itself, how well-educated the networks are and what data is available for that training. When you think of self-driving cars, it is also a version of the machine vision.

They are the mature technology and largely because they have been built by the great technicians, whether they are in the west or in or in the east. When you take it and apply it to financial services and insurance, it is not a surprise to see the things that are most used to being the one who most adapts to human ability.

It makes sense. So how do you see that AI is applied to the traditional insurance chain? In the distribution, for example?

If you are thinking about insurance sales agents, what are they doing? Well, they have a role to be physically present where a client is. You can think of it as almost a sign for the financial product, and in the US I think there are 370,000 insurance sellers – so there is a role for AI there. How do I find the client? How do I get to where they are? Artificial intelligence can help you figure out, based on settings and browsing history, and so on, where your customer segment resides.

The second step from it takes the customer and engages them in some form of conversation. In the physical world you can get a person coming home or going out on a website to make an assessment. In the digital world, the phone is your attention platform.

So it's very important to put your head on because there are only 5 to 10 places on the phone for financial apps. In the physical world you can have as many branches as you like, and you can send as many people as you like – in the mobile world there are only five places you can take. It is extremely important for economic operators to figure out how to live within these attention platforms and have attributes that are built into the attention platforms.

Chatbots are one of those things. (And for me, chatbots and voice are essentially the same.) Chatbots like things that live either inside the phone as a standalone app, or living inside something like Facebook Messenger as a standalone bot. If you like downloading Lemonade or Leo or something similar and being able to communicate in the app, it is just a built-in function of how to build your customer service function. We are in a world where most of the attention is with the major technologies, and not with signs or other types of traditional media advertising. So it is massively important.

Of course, caveat chat bots have not been very effective in replicating a human interaction. It is really difficult to find the line between man and machine, and the negotiation of that line is where you can make or break the customer's experience. If you have a customer coming into your app and trying to discuss something with your chatbot and it is a frustrating experience and they would rather talk to a person you will definitely lose them. And if you don't have an easy way to drive that conversation flow into a human channel, you will only lose that customer.

And then in other cases, as well as by generational lines, you may have a much better experience with the customer who can get on board via the phone, can get access through a phone, can take a picture of the passport to get through Know Your Customer and Anti Money Money Compliance (KYC AML), or can take a picture of the damage to their car and pass it on to the insurance company or for damage assessment.

It is definitely a negotiation between how frustrating it is to work with a chatbot against how nice it is to be able to do these things automatically and quickly. And I think it's still discovered or explored. I would say that we do not have a final answer there – partly because the underlying technology still has a lot of room to go.

Amazon Alexa and Google's AI Assistants are still in their very, very rudimentary steps, and I would expect the next ten years to be these platform shifts where the big companies are competing to talk well. So it is the first paragraph – insurance agents and the role they play.

I would also flag the claim process. About 250,000 people are involved in claims handling, so the size is also quite large. And then if you look at insurance policies and people working on the models, you get about 280,000 people. There are as many opportunities for automation using this technology and all the different parts of the value chain.

I love your description of the phone as an attention platform. Thank you so much for taking the time to talk to us today, Lex.

Summary

In this episode of Accenture Insurance Influencers podcast, we talked about: [19659000] [19659000] 19659040] The difference between automation and AI. Automation is a case of "about this, then it", where the results are well-defined and understood. AI is a probabilistic result from a trained neural network, deployed on a scale, where the results may be unexpected.

  • AI can be used as chatbots for interfaces with customers that insurance agents do today; However, there is work to be done to improve how chatbots replicate human interactions.
  • Claims and insurance guarantees are other points in the insurance life cycle where there may be opportunities to distribute AI.
  • In the digital world, the smartphone is a "real estate" platform with limited property for financial apps. Insurers would be careful to figure out how to live in attention platforms.
  • For more guidance on AI in insurance:

    In the next section, Lex will discuss AI's ethics. How does bias cross into AI-driven decisions and what can insurers do about it? Finally, given the major issues we have covered in this series break, innovation, insurtech and AI, to name a few – what can the existing insurance companies remain competitive without sacrificing shareholder value?

    What to do next: [19659007] Contact us if you want to be a guest at Insurance Influencers podcast.


    Source link