Manulife Deploys AI Agents to Streamline Financial Operations

Manulife, the Canadian multinational insurer, is accelerating its adoption of artificial intelligence by embedding agent-based AI systems directly into its core operational workflows. This strategic transition marks a move beyond limited use cases—such as data analysis or customer service—towards the automation of high-volume business tasks at scale.

The company has developed a new platform that enables 'agentic AI', referring to systems capable of carrying out multi-step tasks across different software environments and data sources. Unlike traditional chatbots, which respond to individual queries, these AI agents can interact with internal databases, process structured data, and execute sequences of actions integral to financial and insurance operations.

According to Manulife, this initiative is part of a wider plan to boost productivity, reduce manual workloads, and facilitate internal decision making. The insurer anticipates generating over US$1 billion in value through AI-driven workflow automation by 2027. Manulife has been gradually increasing its internal use of generative AI—AI systems that can create text, summaries, or reports—with more than 35 use cases currently in production and plans to double that figure in the coming years. The company reports that about 75% of its global workforce is already utilising generative AI tools to some extent.

Insurance companies, which manage vast amounts of structured data such as policy details, claims, and regulatory reports, stand to benefit significantly from process automation. Manulife's new platform allows teams to implement AI agents to navigate internal systems, aggregate information, and prepare documentation for decision-makers. For instance, an AI agent could collect data from various platforms and compile concise reports for staff reviewing insurance claims or financial assessments, thus reducing preparation time.

Across the industry, testing and implementation of generative AI have expanded rapidly. McKinsey's 2024 Global AI Survey found that 65% of organisations now use generative AI in at least one business function, up sharply from the previous year. However, most deployments remain confined to pilot projects or specific departments, with only a minority reaching full-scale production in major operational areas.

For financial institutions, introducing AI into core systems presents unique regulatory and operational challenges. The sector is subject to stringent oversight, demanding strong governance frameworks to ensure fair, transparent, and explainable AI-driven decisions, especially in sensitive areas like underwriting and risk assessment. Manulife states that its platform incorporates comprehensive governance and security measures to monitor decision-making, ensure data compliance, and maintain alignment with internal and external regulations.

The push towards agent-based AI reflects a wider trend among financial firms aiming to automate repetitive, resource-intensive tasks in claims processing, compliance, and reporting. Other banks and insurers in North America and Europe are also piloting AI agents for fraud detection and internal analysis. Research from Accenture suggests AI-driven automation could reduce operational costs for financial institutions by up to 30%, mainly by expediting routine processes and improving accuracy.

Despite these potential benefits, the adoption of operational AI carries risks. Errors made by AI models can be amplified in automated workflows if not carefully monitored, prompting many firms to introduce new systems gradually and focus first on supporting rather than replacing staff.

Manulife's large-scale implementation of agent-based AI signals a key step in the enterprise integration of artificial intelligence. The long-term success of this approach will depend on its ability to improve efficiency while maintaining transparency and regulatory compliance. As financial institutions advance beyond initial AI experimentation, seamless operational integration will be central to realising the technology’s promised value.

AI News