di-zuniga-allstate-iaai.jpg
Allstate SVP Carla Zuniga opened day two of Digital Insurance's Insurance Analytics & AI conference with a discussion of how the carrier is leveraging AI in several areas of the enterprise. Zuniga led the team that integrated IPSoft's Amelia AI technology into Allstate's customer contact center. Amelia has helped customer-service staff stay organized and work efficiently during calls -- "it's not uncommon for an agent to have 13 to 14 items open on their computer while handling a service call," Zuniga notes -- with positive trends in first-call resolutions and call duration metrics. Next is turning Amelia loose with the company's distribution force, giving them the tools to better support Allstate initiatives like Drivewise usage-based insurance.
di-arora-aig-iaai.jpg
Next up was AIG's Neeraj Arora, global head of decision science and data automation for AIG, who presented "four key mindsets for analytical transformation." "God, or science or evolution or whoever, didn't want us to change things fast," he said at the outset of the talk, explaining that getting technology isn't the issue with adapting to new ways of thinking in insurance, but ensuring that staff understand the best ways to leverage that tech can be.

The four mindsets are:
1. What to ask. "Changing perspectives reveals hidden truth," Arora said, noting that different metrics can illustrate different kinds of successes with programs and it's important to look at data from other angles.
2. What data to use: "It's hard to clean data or find it without bias, but that can be identified," Arora said, using telematics as an example: Does telematics actually reduce losses, or do better drivers simply sign up for UBI because they know they will be successful. It's important to have a control group to measure results against, he says.
3. What methodology to follow. When faced with a new task, there are several ways of tackling: One could modify processes, introduce extra technology or retrain existing staff. "A mix of these is most successful," Arora says.
4. Who to involve and how. "Analytics can't be scaled if limited to one team, the entire organization needs to embed an analytical mindset in day to day work," Arora says. That means creating solid illustrations and dashboards so that all users can visualize the impact of data and analytics programs.