writes, “The global health crisis has inevitably impacted the insurance industry, with some experts predicting that 2020 will be one of the most expensive years the industry has ever experienced.
However, in the motor insurance sector, this is offset by the reduced number of claims, as people stay at home and cars sit idle. British start-up, Cuvva, estimates that there will be a 50% reduction in car insurance claims, amounting to a £1billion windfall for insurers.
Driving customer loyalty
Some insurers jumped on the opportunity to win customer loyalty by repaying them a portion of their premiums for the weeks when vehicles weren’t being driven. Admiral made the first move in the UK market, announcing that it had set aside £100million to repay £25 to each of its 4.4 million motor insurance customers. In the US, AllState said it would repay 15% of customers’ premiums for April and May. A day later, Geico announced that it had committed $2.5billion to repay dividends to customers whose vehicles have not moved during the lockdown period. American Family also announced plans to reimburse customers $50 per insured vehicle.
Looking after customers
When lockdown measures were first introduced in March, the immediate impact was on customer interactions. As social distancing measures were introduced, some insurance companies found that contact centre staff overseas did not have the connectivity to be able to handle calls from home. This resulted in fewer employees being available to answer calls. With a reduced workforce handling higher call volumes, customers initially faced long wait times and increased resolution times, resulting in a poor experience.
However, the insurance industry is swift to respond to challenges and has been steadily implementing technology to drive out cost, while improving customer experience.
Applying AI to drive down costs
Three years ago, Accenture’s ‘Technology for People’ report, found that 68 percent of the surveyed insurance professionals said that they were already using AI-powered intelligent virtual assistants to improve customer interactions. More than three quarters (79 percent) told Accenture that AI would be used to revolutionise the way that insurers gain information from and interact with their customers.
So, it’s not surprising that, as contact centres experienced a sudden spike in call volumes handled by fewer customer service agents, insurers were able to swiftly implement AI to handle enquiries, in parallel with human agents, so that customers did not feel abandoned.
Setting up a chatbot in 48 hours
One such example is AA Ireland, which set up its COVID-19 chatbot within 48 hours. In the first week of March travel restrictions were announced in Ireland and the contact centre in Dublin began seeing a rapid increase in the number of calls from travel insurance customers. By 9th March, its COVID-19 chatbot was launched via its IVR menu.
The insurer was able to use the AI-based virtual assistant to rapidly follow a three step process to first deflect callers from phone lines to find the information they needed on text-based channels such as its website, mobile app, messaging solutions and social media. Secondly, by sending the customer a link to a virtual assistant, the AA Ireland could automate conversations and interactions while making customers fully aware that they were getting information from a virtual assistant. Thirdly, the insurer integrated the chatbot with its existing Zendesk live chat platform, so that a customer simply had to type, ‘I want to speak to someone,’ for the bot to hand over the conversation to a human agent.
Customers receive an SMS on their mobile with a link to the virtual assistant that instantly pops up a COVID-19 message pointing to related updates. Since the virtual assistant also has capabilities around other common customer service queries, it offers them multiple options besides COVID-19 information updates, for example, requesting documents, renewal assistance, or making policy changes. This allows customers to self-serve in many ways and to dip in and out of the process, going back in to the COVID-19 bot to get information when it’s convenient for them.
After the COVID-19 chatbot went live, the contact centre saw an 11% reduction in calls as customers were directed to the information that they needed. The virtual assistant integrates with the livechat system to enable seamless handover and escalation to live agents.
Steps to success
Insurers don’t need coding knowledge to be able to benefit from conversational AI platforms. They need an understanding of customer processes, which is where their contact centre professionals excel.
Natural language processing (NLP) and AI bot technologies are advancing at a rapid rate. With an abundance of training, combined with the advancements in natural language AI models, a virtual assistant can carry on an impressive and lengthy conversation with a customer: recognising context, sentiment, and nuances of language.
It’s vital to make customers aware at the outset that they are conversing with a virtual assistant, rather than a livechat agent.
To mitigate any impact on the customer experience, seamless integration with livechat is required, so that customers don’t get caught in a chatbot loop and can always speak to a customer service agent if they need to.
Post-contact analysis should be carried out to quantify how many virtual assistant conversations have escalated to a live agent, so that conversations can be improved. A/B message testing should also be used to optimise the AI bot, in the same way that customer service agents tailor conversations to suit customers’ requirements while they’re on the phone with them.
Even when the conversations have been perfected, to really drive down operational costs, insurance AI assistants also need to be able to automate underlying workflows and processes. No matter how good a conversation is with an insurance bot, there will quickly come a point where customers need to input their payment details. Security has to be prioritised for any insurer that wants to create an enterprise-grade conversational AI experience.
Each industry and each company has unique requirements. The insurance chatbot will need good initial training data, based on realistic customer conversations and the AI model will need to be tuned and optimised on an ongoing basis.
Staying close to customers
Many companies are advancing their AI journeys, launching virtual assistant solutions that are highly skilled in a specific business task, whether that is helping a customer with a quotation, onboarding new users, or reminding customers that their policy is nearing its renewal date. Virtual assistants are highly successful at handling these customer interactions: reducing agent handling time, driving increased conversion rates, and lowering the cost of service delivery.
As virtual assistant implementations become more sophisticated, natural language technology will allow companies to understand more precisely what customers want, so that automated processes meet the customer need, reducing manual intervention. In the meantime, it is vital to enable virtual assistants to handover to a customer service agent if a customer requests it, or if a bot hasn’t been trained to respond to a particular query.
During the COVID-19 pandemic, this combination of AI-based automation and human agents enabled insurers to rapidly scale their customer service operations, while keeping costs in check. As the insurance industry navigates the post-pandemic era, it will need to seek more ways to apply AI to reduce operational costs, while keeping customers on board.”