This post is also available in: Français (French)
Insurance Nexus publishes a report titled “What is the true value of automation?” intelligent? ” in support of their two fall events: Insurance AI & amp; Analytics and Connected Claims Europe Summit .
Note: This article is written as part of a media partnership between my blog and Insurance Nexus to promote their publications and events. Nevertheless, I am not compelled to write a favorable opinion of this report. Anyway, I think it’s relevant to my audience, just like the previous reports that I relayed and that you can find here or here for example.
Insurance Nexus shares the view, as I do, that there are many buzzwords used around the world for insurance innovation. They wanted to truly realize what is happening today around artificial intelligence. So they conducted a dozen interviews so the white paper so it is question here is the synthesis.
Here are the key points!
What is the reality of automation today?
- The best customer relationship is the one that manages stress for the customer. Automation aims to improve process efficiency to satisfy this customer relationship.
- Automation is therefore about replicating the existing, better.
- Robotization can be used in many cases of use to avoid repetitive operations. For Steve Tait, it is the claims that have the most to gain from automation.
- This is obviously not limiting because automation can be a good way to connect very disparate data from multiple “legacy” systems. This makes it possible to obtain fast results without too much questioning.
- Mistrust all the same, automation or robotization is not an end in itself. These are ways to progress. The goal is to have structured data of good quality and stored in a sustainable way. However, it is complicated to realize. The intermediate step of automation is therefore very useful.
Manage internal expectations
When we talk about automation, people often think that it’s about asking the machines to solve our own problems or difficulties. This is not the case, the machines do not solve our problems, they transform them! You have to understand what artificial intelligence is capable of before you start asking anything.
Thus, robotizing is one thing, but to move towards automation intelligence, it is about looking at the processes, from the extraction of data, its interpretation and the decision making associated.
In the last 5 years, there have been many projects around artificial intelligence that lacked preparation. More specifically, it is the design phase that is left out. People do not know what problem they are trying to solve! When you buy a technology, you buy a solution, but if you do not know what problem this solution meets, it is useless.
Finally, it is important to remember that automation will not replace jobs, but it will transform them. The role of sinister manager for example will not disappear, even on a horizon of 10 years. On the other hand, it will dramatically change.
Artificial intelligence is going to have more and more impact
If those who are convinced of the interest of artificial intelligence are numerous, the craze is not yet quite there.
For the time being, work is mainly focused on identifying relevant use cases that could merit automation. Thus, it is above all a process of continuous improvement which is announced, gradually progressing towards more and more complex subjects
Chatbots are settling down to last
Chatbots thus appear as an interesting interface to the use of artificial intelligence as a vector of intelligent automation. Indeed, thanks to the use of Natural Language Processing (NLP) technology, it is possible to consider simplifying the life of the customer by making available what he seeks by sparing him a tedious search.
These solutions also make it possible to consider bringing personalized information where it is needed. For example, during a disaster, it is possible to inform the driver of the status of his request, ask him where he is, the type of vehicle and accident involved, and inform him of the delay. waiting for the tow truck. This raises quite broad questions about the exploitation of the data at his disposal, in an automated way, but especially intelligent and in the sense of the answer to the wait (expressed or not) of the customer.
Significant productivity gains
The gains expected from automation are not negligible. For example, the use of robots allowed a claims processing unit to process data in 30s where it took 10 minutes before. Multiply that by the volume and it opens interesting horizons.
In the same way, payments to suppliers, by their number and complexity, can be automated, with a standardization phase.
In short, on intelligent automation
This white paper is only a preview of what will be proposed during the two Nexus events. However, it already raises quite a lot of relevant ideas about the use to be made of artificial intelligence.