DWIC18 - Insurers must cut down AI strategies to less than 12-months to avoid implementation failure

Insurers must adopt a ‘rapid response’ to implementing artificial intelligence within their business to gain a competitive edge and unlock the financial potential presented by AI.

This was the advice given by Satadru Sengupta, general manager and data scientist, insurance, at DataRobot, on the second day of the Dubai World Insurance Congress 2018.

Sengupta defined the timeframe appropriate of a ‘rapid response strategy’ as less than 12-months. Anything beyond this was described as a ‘long-term’ strategy, which historically contributed to implementation failure.

He said: “Don’t create a two-year or even a one-year implementation strategy. Try to make it a three- month project.

“New technologies such as data robotics are extremely efficient and highly automated, and because of this, they are capable of generating valuable data in only six minutes.”

Companies that adopt a ‘long-term’ strategy will only realise failures in the implementation process six- to 12-months down the line, at which stage, this can be considered as a “point of no return”, he warned.

“What we really want to do is discover the failures earlier than this”, he said, to ensure problems can be addressed efficiently and that AI plans are embedded within the business successfully.

Sengupta rounded-up his recommendations with two more practical tips:

1. Create an education programme on the core benefits of AI for employees at all levels. The education programme must embrace change management and any perception that AI is a cost should be challenged and countered by metrics that can demonstrate potential return on investment.

2. Don’t get too hung-up on the challenges of analysing high volumes big data. Sengupta argued that AI could help extrapolate value, even from the smallest amount of data.