Insurtech Top 5  – DIA Munich and AI Data Analysis

November 22, 2017

Artificial spent a couple of cold days in Munich last week for the third iteration of the Digital Insurance Agenda conference. Startups from all over the world, but mainly Germany and Holland, demoed their technology on stage (no slides allowed), although take note, trying to talk to your virtual insurance assistant does not work in a crowded hall. These were our top takeaways:

1 — Claim Automation

A lot of the startups were offering insurance companies the chance to improve their claim customer experience. Keylane provide insurance companies with the opportunity to improve the first notice of loss process and remove the standard 5–10 day claim time lag. They also integrate with Friss to detect a gadget from a photo, which automatically carries out a fraud check. The platform then goes on to extract information from invoices as well as carrying out value checks across different retailer websites, so you know how much you should be reimbursed. Clever.

2 — Data

Not surprisingly, a lot of insurtechs aim to help insurance companies with data. There were different examples of how this can be done. By improving efficiency, like Squirro, who have an unsupervised learning AI, that analyses the massive amount (40 zettabytes) of data that insurers are unable to doing anything with, automating processes and connecting to reinsurer’s policy admin systems. Or, like Understory, who use weather data to help insurers make decisions, both for rating purposes as well as post claim valuation purposes. There is a huge amount of damage from hail storms in the US and a lot of the claims that are submitted are dubious. Understory allow insurers to track the route of the storm to analyse whether a claim is fraudulent or not, saving an adjuster’s wasted journey to the property.

3— Chatbots

A key feature of a lot of the startup’s tech were chatbots. Regardless of whether you are getting a quote, amending your policy or making a claim, you will be chatting to a bot the vast majority of the time very soon. For example, One, which was recently acquired by WeFox, offers you the chance to manage your insurance from an app, through a chatbot. The aim is for 60% of claims to be dealt with through the app. Customers obviously still want to talk to someone over the phone if they have a large loss, but chatbots are on the rise.

4 — Health Tech

Whilst there were only a handful of startups concentrating on health, they were definitely among the most interesting. Skinvision is trying to tackle the huge problem of skin cancer using machine learning, expert dermatologist advice and everyone having a smartphones. Users take photos of their moles and the app will analyse the photo and tell you whether the mole is dangerous or indeed, cancerous. The bigger their database, the better their algorithm will be able to diagnose cancerous moles.

Another interesting health orientated insurtech are Allm, who have created a Whatsapp for doctors dealing with patients with acute diseases, mainly strokes. $700bn is spent every year on stroke care and the main reason for this is the delay in the time taken to diagnose which part of the brain the stroke has occurred, leaving victims with rest of life care. Doctors will know when a patient has arrived in the hospital, what care they are receiving and are able to consult over the app — reducing ‘door to needle’ time.

5 — Cars

Another theme was the number of startups aiming to help insurers with car claims. Tractable seem to set the benchmark with their AI engine analysing photos of car claims, matching it with the make and model, advising whether the damage is repairable or not and then telling you how much the repair costs would be. Time and money saved.

By Max Odlum, editor of artificial. blog