Jakub Jirsk - Fotolia
How Manulife is empowering sales and call centre agents with AI
Canadian insurance giant is leveraging generative AI to personalise customer experiences, transform contact centre workflows and automate underwriting processes
Canadian insurer Manulife has been on a billion-dollar digital transformation journey over the past few years, leveraging the power of artificial intelligence (AI) to streamline its operations and enhance customer experiences.
In an interview with Computer Weekly, Mark Czajkowski, chief analytics officer for Asia and chief marketing officer for Singapore at Manulife, stressed the importance of having a robust data infrastructure that has enabled the company to deploy hundreds of machine learning models in production.
“The foundations need to be well established and structured in the right way,” he said. “Because ultimately, once you start plugging in all these models, if your data is of bad quality or not captured in the right structure, then you’re hamstrung in what you can do.”
The explosion of generative AI (GenAI) has further accelerated Manulife’s AI initiatives. “The attention has triggered a lot of innovation and curiosity from the senior leadership team,” said Czajkowski, noting that this has driven the development of new infrastructure to support real-time, application programming interface (API)-driven GenAI applications.
Manulife’s AI strategy is centred around key use cases that benefit multiple markets. One primary focus is sales agent enablement. As agents’ customer portfolios grow, it has become increasingly difficult for them to keep track of individual needs and preferences.
To address this, Manulife has integrated AI-powered insights directly into its agent platform. “For each product category, we can tell the agent very quickly what the current coverage and potential gap is,” said Czajkowski.
The platform leverages GenAI to generate engagement ideas and talking points for each customer, enabling agents to have more effective conversations and tailor their recommendations based on individual circumstances.
Machine learning
But it’s not all GenAI. Underpinning those recommendations are machine learning outputs that have to be transformed before they can be leveraged by GenAI, said Czajkowski.
“For example, a machine learning model may generate a score of 0.98, but GenAI will not know what that means,” he said. “We have to refactor the output, such that a score of between 0.6 and 0.98 means you have a higher propensity for retirement needs. We feed that into prompts which then get interpreted.”
Another application of AI is in Manulife’s call centres where agents spend a lot of time looking up contract documents and consolidating information. Manulife is using GenAI to automate this process, allowing call centre agents to access relevant information quickly and provide faster service.
Czajkowski said that while the sales agent platform leverages machine learning models, other capabilities such as contract lookups for call centre agents primarily use retrieval-augmented generation (RAG) techniques to consolidate information efficiently.
Addressing the laborious underwriting process, Manulife is also using GenAI to ingest and analyse medical documents, summarising key information from various sources for easy review.
“We’ve been able to automate all that work for underwriters and structure the information in a way that’s easy for them to understand, so they can spend more time doing adjudication and decision-making work,” said Czajkowski.
The underwriting application also flags potential inaccuracies in transcribed medical data due to handwriting or other issues, allowing underwriters to review and correct the information, which is then fed back to improve the model’s accuracy. An AI chat assistant is also on hand to help underwriters quickly access specific information in the documents.
Reiterating the importance of data quality, Czajkowski acknowledged that GenAI is not “magic”, noting that “the effort really needs to go into the data, and how it’s structured and transformed”.
This data-centric approach is reflected in Manulife’s revamped system development process, with increased emphasis on capturing and managing data quality upfront. Cross-functional squads comprising IT, data science, as well as user experience (UX) and user interface (UI) professionals work collaboratively to deploy AI use cases, fostering a shared understanding of data’s importance across the organisation.
Manulife is also taking steps to ensure the responsible use of AI. The company has established a model governance process that includes materiality assessment of use cases, model risk management, and alignment with ethical principles and regulatory guidelines. Czajkowski said this process ensures responsible data usage, manages data drift and guides model retraining efforts.
While acknowledging the high cost of implementing GenAI technologies today, he said “it’s getting cheaper, which means more use cases can be enabled”, adding that Manulife evaluates the business case for each AI initiative to ensure the benefits outweigh the costs. “It’s something we look at naturally when we start investing in any project,” said Czajkowski.
Read more about AI in APAC
- Malaysia’s UEM Sunrise is leveraging Salesforce and AI to boost sales and enhance customer experience while prioritising data sovereignty and a diversified cloud strategy.
- Vietnam’s Techcombank is targeting $1bn in profit in 2024 while maintaining a flat headcount and reducing physical branches, thanks to its investments in AI and a data engine.
- Nvidia is expanding its presence in India through new partnerships focused on building AI infrastructure, developing Indian language models and training AI professionals.
- At the Gartner IT Symposium 2024 in Gold Coast, experts call for CIOs to pace their adoption of AI, manage costs and embrace ‘augmented leadership’.
