(Image source: Manulife.)
Manulife Financial Corporation (Toronto), one of the world’s largest life insurers—with roughly 37,000 employees and more than C$1 trillion in assets under management and administration in their Manulife Wealth and Asset Management business—is moving to industrialize its artificial intelligence capabilities as part of a broader strategy to become what executives describe as an AI-powered insurer.
The initiative accompanies the insurer’s selection of Akka (San Francisco) to provide runtime infrastructure for a new enterprise AI platform, alongside AdaptiveML as its reinforcement-learning engine to dynamically optimize growth and govern AI use cases at scale. According to Jodie Wallis, Global Chief AI Officer, Manulife Financial Corporation, the effort reflects a fundamental shift in how the company views artificial intelligence – moving from isolated experimentation to a disciplined value-driven approach that embeds AI responsibly across the enterprise.
For the insurance industry, Manulife’s move highlights how rapidly AI is shifting from experimentation to enterprise infrastructure. Many carriers have launched AI pilots, but far fewer have begun building the governance, compute management and operational platforms required to run hundreds of AI applications inside core workflows. As one of the largest global life insurers and a global asset manager, Manulife’s strategy suggests that the next phase of AI adoption in insurance will be defined less by individual use cases than by the ability to operate AI as a foundational capability across the enterprise.
Jodie Wallis, Global Chief AI Officer, Manulife Financial Corporation. (Click to enlarge.)
“In our refreshed enterprise strategy, driving AI value is one of our five key pillars,” Wallis says. “Before 2025 it was an enabler of business strategy, now we believe it is an integral part of the business strategy.”
That change reflects the magnitude of AI’s potential impact on the insurer’s operations. Wallis describes AI as a technology capable of reshaping nearly every aspect of the business—from customer experience and distribution to underwriting and fraud detection.
“AI has the potential to impact everything we do.,” she comments. “AI is something that will change all parts of our business, it’s good for our customers, colleagues and stakeholders.”
A Decade of AI Investment
Manulife’s AI program dates back nearly a decade. The company began investing in artificial intelligence capabilities in 2016, initially focused on machine-learning applications such as propensity modeling, pricing analytics and fraud detection.
For several years those initiatives largely supported analytical decision-making rather than operational processes. Between 2016 and 2023 the insurer deployed roughly 70 AI use cases.
The arrival of generative AI accelerated development dramatically.
“In 2024 and 2025 we deployed 140 AI use cases,” Wallis says. “That’s roughly 70 a year. In 2026 we plan to deploy another 200.”
The change represents both a sharp increase in deployment velocity and a shift in how AI is used. Rather than operating primarily as offline analytical tools, AI systems are increasingly embedded directly into operational workflows.
“As AI becomes part of high-volume business processes, it becomes a core operational capability,” Wallis explains.
From Experiments to Infrastructure
The decision to build an enterprise AI platform—and select Akka to support it—emerged from Manulife’s effort to prepare for that rapid scaling.
Before selecting a technology partner, the company first asked what capabilities such a platform would require.
Wallis identifies three priorities that shaped the platform’s design.
The first was the need for a standardized developer experience. As AI development expands across the organization, Manulife wants to ensure consistent engineering practices and governance controls.
“We’re moving from a small group of specialists deploying AI solutions to a wider group of colleagues building solutions,” she says. “We need a consistent developer experience that will support the development of reusable, high quality and responsible AI solutions.”
The second requirement reflects the operational nature of modern AI systems. Many applications now run directly inside customer-facing and internal workflows, creating new requirements for reliability and responsiveness.
“If AI is going to be part of how we deliver services to our customers and employees, then we need to treat it like any other high-availability platform,” Wallis says. “That means low latency, high availability and full visibility into what’s happening across those systems.”
The third factor was compute efficiency.
Large language models and other advanced AI systems rely heavily on GPU-based computing infrastructure, which has become both expensive and scarce.
“For many years we assumed compute costs would keep going down,” Wallis notes. “With LLMs, we’ve hit a point where that assumption no longer holds.”
Manulife therefore sought a platform capable of optimizing compute usage while balancing performance, financial cost and environmental considerations. The company already operates one of the more cloud-centric IT environments in financial services, with more than 80 percent of applications running in the cloud.
Responsible AI at Scale
Rapid deployment does not come without safeguards. Wallis stresses that Manulife is expanding governance capabilities alongside AI development. Manulife’s Responsible AI Principles empower the firm to deliver value from AI for its customers, colleagues and society.
All AI solutions are evaluated under the company’s model risk management framework, which assesses systems based on their materiality and determines testing and validation requirements.
Manulife’s AI Principles. Source: Manulife. (Click to enlarge.)
“We’re not skimping on governance,” Wallis says. “For every dollar we are investing in deploying AI solutions, we are also investing in AI safety.”
The company also employs a range of technical techniques to improve reliability and reduce model failure risk in generative AI systems.
One such method uses one model to generate an answer and another to validate it. Another approach involves limiting AI systems to specific corporate knowledge sources rather than allowing unrestricted access to external data.
“We might tell the model to ignore everything it has learned except for a defined set of documents,” Wallis explains.
Manulife also conducts adversarial testing, running systems against known prompt attacks and misuse scenarios collected in open-source repositories.
Priority Use Cases
Wallis highlights several areas where AI is expected to deliver the greatest impact.
Distribution is one of them. In Asia alone, Manulife works with roughly 106,000 agents and advisors.
AI tools can help those agents prepare for meetings, understand customer needs and tailor communications.
Underwriting represents another opportunity.
Life insurance underwriting often involves large volumes of unstructured information, including physician statements, lab results and other medical records. Before any underwriting analysis can occur, much of that information must first be organized. “With AI we can extract what’s relevant and structure the information much more efficiently,” Wallis says.
Manulife was the first life insurer in Canada to use AI in underwriting. The firm recently introduced a redesigned electronic application and an enhanced version of its proprietary AI underwriting engine, MAUDE (Manulife Automated Underwriting Decision Engine).
Similar capabilities are emerging in claims processing, particularly in markets where Manulife offers health insurance products. AI can analyze non-standard claims documents and extract relevant information for automated adjudication.
More broadly, Wallis describes a growing category of applications focused on intelligent document processing.
“Give me any document anywhere, anytime in any format,” she says. “With AI, I can make sense of it, extract what’s important, and allow the automated part of the process to proceed.”
AI for the Workforce
Manulife is also deploying AI tools internally.
These include virtual assistants that help employees handle tasks such as HR inquiries, procurement requests and IT support. The company is also using AI to assist software engineers with code development, testing and environment provisioning.
Wallis refers to this category of applications as “AI for tech.”
Encouraging employees to adopt AI tools has also been a priority. Manulife has introduced workshops known as “prompt-a-thons,” where teams examine their daily workflows and experiment with AI prompts designed to improve productivity.
“We sit down with people and ask: what’s on your calendar today? What’s on your to-do list?” Wallis explains. “Then we explore how AI might help them complete those tasks more easily.”
Those initiatives have contributed to widespread adoption across the organization. In 2026, more than 70 percent of Manulife’s workforce now uses AI tools regularly.
From Data-Driven to AI-Powered
As insurers have spent the past decade working to become data-driven organizations, the next phase may involve operationalizing that data through AI embedded in core processes.
Wallis suggests that shift is already underway at Manulife, where AI is moving beyond analytics into underwriting workflows, customer support and internal operations.
“In the past we talked about being data-driven,” she says. “Today we talk about being AI-powered.”