Agent-Facing AI For Swifter Insurance Support

In the aftermath of natural disasters like Hurricane Florence, insurance call centers must also rush to accommodate the surge of incoming calls. In the latest Call Center Tracker, Allstate SVP Carla Zuniga explains how AI can help agents resolve complicated queries and speed up onboarding, while customer-facing automated tools can route calls and resolve basic requests — all for faster support.

Natural disasters have shaken communities across the world this month, from Typhoon Mangkhut raging in the Asian-Pacific to Hurricane Florence battering the U.S.’s eastern coast. The recovery process after any catastrophic event often requires people to reach out to their insurance companies for funds to help clear the path to stability and normalcy. Those affected need to call contact centers to file a claim and get help, or go through other channels such as websites, mobile apps or visiting a staffed office location. Even in less drastic situations such as automobile collisions or more routine property damage, contacting insurance companies is key to restoring order. In a recent interview with PYMNTS, Carla Zuniga, senior vice president of Allstate, discussed the strategies the company uses in its call centers to provide quick support to customers during times of high demand.

Faster onboarding

To prepare its call centers for all manners of customer needs, Allstate has turned to an agent-facing AI-powered virtual assistant to get new hires up to speed quickly. The solution, Amelia, advises agents during customer calls by providing information on legal restrictions, policies, and other guidance depending on the customer’s request. That information can be particularly important in a highly regulated industry.

“[The agents] are guided through the call — not necessarily with a script, but rather with the rules and guidelines that apply to their role,” she explained. “[Amelia helps] them understand what they’re empowered to do for different customer inquiry types, [what they] might be limited to do from a licensing standpoint, and the policies and procedures that apply to those requests.”

Quicker onboarding is especially valuable in call centers, where retention can be a challenge, Zuniga said. It’s even more important for the insurance agency, where onboarding often is a more lengthy undertaking due to the number of regulations that agents must be familiar with. Their training can be an 8- to 12-week process, but Amelia helps shave a few days off of the curriculum, allowing agents to focus on more common scenarios, since they can fall back on Amelia for support in rarer cases. Agents usually take 20 to 24 weeks to reach a skill level regarded as “proficient,” but with Amelia’s support, they’re able to provide proficient-level service four weeks sooner.

Allstate also uses automated tools help employees from outside the main claims handling team refresh their training, enabling them to pitch in when there’s an influx of demand. The technology enables the company to enlist dozens or hundreds of representatives who don’t normally spend their days on customer phone calls, helping them with first call responses and initial documentation, before they pass customers on to claims specialists.

“Technology does support some of the quick and easy onboarding of representatives who may not spend a lot of time on the phone … but who are some of our residual support we depend on,” she said. “We keep training fairly simple and straightforward so they’re ready when customers call.”

Automation and self-service

On the other hand, forthcoming customer-facing automated systems are intended to handle initial inquires and loss procedures, while directing only those that need it to a live agent. That frees specialists to focus on more complicated cases.

“Automation helps flatten out the demand curve,” Zuniga said. “[It] is a key strategy in being able to respond to catastrophic situations and general seasonal fluctuations that we see.”

The company has been testing natural language chatbots for routing and resolving customer requests. It is also considering bringing virtual assistant capabilities into voice support, text and chat interactions. Customers have shown positive responses to self-service options, readily turning to the insurance company’s mobile and web channels to conduct straightforward transactions that don’t require speaking with a professional. Looking forward, Zuniga expects the use of automation to increase.

“Eventually, [with expected developments,] the need to talk to human representatives will be minimized,” she predicted. “We’ll save those interactions for the more complex, highly-regulated, difficult transactions that we’d never [force] customers to solve by themselves with a machine.”

In the coming years, Zuniga expects there will be deeper levels of machine learning and systems that are capable of identifying patterns and leveraging that information to proactively serve customers. Even as the use of digital channels and automation continues to grow, she said technology will not replace the need for human representatives, but instead bolster their efforts.

“Technology can supplement the relationship and support the representatives that are working largely in case management and complex insurance situations, as well as advise and guide our agency force through the very best ways to focus [and grow] the relationship …and increase protection,” she said. “It’s figuring out how great technologies can augment the relationship-based organization that we are, whether that’s internal or with customers.”

In times of instability, sometimes the most reassuring thing is having a person there and ready to help. Allstate’s support even includes bringing staff to catastrophe areas in mobile units to act as local, on-the-ground service hubs, providing face-to-face assistance, alongside agents that are available on the phone. Automation ensures those live agents aren’t tied up handling basic inquiries and procedures, but instead are ready to roll up their sleeves and dive into the complex work of helping customers get back on their feet.