Lifting the lid on our tech
However, the challenge is also more complex. While cars are relatively uniform – and some models are widely used around the world – houses don’t come in set models or specific dimensions. So how does the technology work?
We started by breaking down the questions the AI has to answer so they become quite narrow. We might ask what material a damaged object is made of, what its dimensions are, and whether the damage can be repaired. Once we have extracted this information, we can use expert rules that we’ve developed to bring everything together in a cost estimate.
For example, with a damaged window, AI Property would extract the dimensions, the glass type and frame material from the image. Then after analysing the image and comparing it to others in its database, as well as local pricing guidelines and standards, it would say, “This is repairable: and for this damage, we need these materials and one person working for a half a day. In this region, it will cost X.” The claim is then approved and paid.
The reduction in cycle time is the most obvious benefit. But the overall processing cost is lower too. The estimate is more reliable and consistent (as contractor quotes for jobs may vary). Companies can better manage large numbers of assessors for short bursts of work and, by having AI deal with the bulk of claims, specialist staff are freed up to deal with more complex ones where they can make the biggest difference.
All this feeds through to savings for customers but perhaps most importantly, it means customers are happier. They have a better, less stressful experience during what can be a very difficult and upsetting time.