Manulife has elevated AI into a pillar of its enterprise strategy and expects the technology to deliver more than $1 billion in enterprise value by the end of 2027, with roughly 30 percent of that target already realized by late 2025. The Toronto-based insurer is moving AI beyond individual use cases and productivity tools, treating it as an expectation for how work gets done across the organization.
"We now look at AI less as an enabler and more of an expectation for how we get work done across the organization," said Anna Havens, Global Head of AI Strategy at Manulife. Havens said the shift became formal late last year when AI was elevated to a pillar of the company's enterprise strategy.
A common platform, not isolated projects
Manulife has organized its enterprise AI work around a shared platform designed to provide common standards, embedded governance, and built-in guardrails. Before this, separate development environments across teams and markets had made it difficult to scale and share capabilities.
"This common foundation allows us to build common capabilities with a common set of standards, including embedded AI governance and guardrails," Havens said. "This is really going to be the foundation for acceleration and reuse across the organization."
The insurer has grouped its AI efforts into six focus areas: virtual assistance, underwriting, distribution, advice, AI for developer efficiency, and AI automation. Each area is treated as a product domain with a global roadmap rather than a collection of one-off deployments. For organizations operating in the AI for Insurance space, Manulife's approach signals a departure from pilot-heavy strategies toward operational integration.
Partner-agnostic by design
Manulife's platform is deliberately partner-agnostic, a choice Havens said helps the insurer avoid overdependence on any single provider as AI capabilities evolve. The company buys for speed and builds for differentiation it cannot get from the market.
Havens identified AdaptiveML as a partner helping Manulife fine-tune small language models, noting that not every AI problem requires a large language model. She also named Akka as the orchestration layer supporting consistent delivery at scale. The broader platform, she said, combines centralized capabilities with localized teams embedded in the business who understand specific workflows and customers.
A finance-validated value claim
Manulife measures AI value across four categories: cost savings, cost avoidance, revenue generation, and fraud prevention. Every benefit must be formally validated by both the business unit and finance before it counts toward the $1 billion target.
"Our segment CFOs govern that process specifically to confirm that benefits are real, to make sure that things aren't double counted, and that all of our business cases are built on really sound assumptions," Havens said.
The company had delivered about 30 percent of the target by the end of 2025 and said it remains on track through 2027. This discipline around value measurement reflects a broader shift among companies featured on AI for Executives & Strategy resources, where finance governance is becoming a standard expectation for enterprise AI programs.
AI fluency as an adoption accelerator
Havens described foundational AI fluency across the workforce as an essential accelerator. Employees need to understand how to use AI in their daily work before specialized, domain-specific tools can scale effectively.
"If we don't have an employee base that understands the basics of how to use AI in their day-to-day, when we go to release domain-specific capabilities for them, it's going to be much harder for them to adopt those," she said.
Manulife is working with HR and learning teams to identify the skills employees will need. Havens pointed to critical thinking, judgment, creativity, and the ability to ask better questions. She noted that AI can be overly affirming - "it's not naturally critical in its responses to you" - which makes independent judgment more important, not less. She also highlighted translation as a key leadership skill, describing her own role in explaining complex AI concepts in language the broader organization can understand.
Responsible AI from the starting gate
Responsible AI is built into Manulife's development process, not layered on afterward. Use cases undergo risk assessment before development begins. Lower-risk internal tools move more quickly through automated governance, while higher-risk customer-facing tools receive full human review.
"Nothing just goes live," Havens said. Human accountability remains a core principle, though it means different things depending on the use case. In underwriting, the company is comfortable using AI for low-risk automated assessments, freeing capacity for more complex evaluations where human judgment remains essential.
Why this matters for executives and strategy professionals
Manulife's approach offers a concrete model for shifting AI from a collection of experiments to a managed enterprise capability. Three elements stand out: a shared platform that enables reuse across markets, finance-governed value measurement that distinguishes projected benefits from realized ones, and fluency-building that treats adoption as part of the operating model rather than an afterthought. The $1 billion target attached to a formal validation process also gives executives a benchmark for tying AI strategy to measurable business outcomes - not just activity.
Your membership also unlocks:
