Large financial firms have spent years testing artificial intelligence in small projects, often limited to data analysis or customer support tools. The next phase appears to involve something more operational: systems that can take action across business workflows. Canadian insurer Manulife is moving in that direction as it works to deploy what it calls agent-based AI systems inside its internal operations.
The company is building these capabilities with a runtime platform designed to support “agentic AI,” a type of system that can carry out tasks across different software tools and datasets. Manulife said the effort is part of a broader plan to automate high-volume work and assist internal decision making across the business.
In a company statement announcing the project, Manulife said it expects artificial intelligence initiatives to generate more than US$1 billion in value by 2027 through productivity gains, improved decision support, and workflow automation. The insurer has been investing in AI for several years, but the current push focuses on integrating the technology more deeply into day-to-day operations rather than running isolated experiments.
Manulife has already been expanding its internal use of generative AI tools. The company said it currently has more than 35 generative AI use cases in production and plans to expand that number to about 70 in the coming years. It also reported that around 75% of its global workforce already uses generative AI tools in some form, according to company disclosures.
Moving AI from experiments to operations
Insurance companies handle large amounts of structured data. Policy information, claims records, underwriting assessments, and financial reports often move through several systems and teams before a decision is made. These processes create an environment where automation tools can assist with tasks such as document review, claims processing support, and internal reporting.
Manulife said the new platform will allow teams to deploy AI agents that can interact with internal systems and data. Instead of responding to a single prompt like a chatbot, these agents are designed to complete sequences of tasks across different software tools and workflows.
For example, an AI agent might collect data from several internal systems, organise the information, and prepare summaries for employees who are reviewing cases or preparing reports. The goal is to reduce the time staff spend gathering information before making a decision.
The shift reflects a broader trend in enterprise AI. Over the past two years, many companies experimented with generative AI tools for tasks like writing, coding, or summarising documents. Analysts say the next challenge is turning those capabilities into systems that can support operational work across large organisations.
A report from McKinsey’s 2024 Global AI Survey found that about 65% of organisations say they now use generative AI in at least one business function, up from about one-third in the previous year. However, the same research notes that only a small portion of those deployments have reached full production across large parts of the business. Many still remain limited to pilot projects or specific teams.
AI inside regulated financial systems
Financial institutions face extra hurdles when they try to move AI into production. The sector operates under strict regulatory oversight, which requires strong controls around data use, model behaviour, and decision transparency. Systems used for underwriting, risk analysis, or investment decisions must be auditable and explainable.
That environment makes governance and monitoring central to any AI deployment. A study from Deloitte on AI in financial services notes that banks and insurers are increasing investment in model oversight tools, internal AI policies, and risk review processes as they expand automation. Organisations are trying to balance efficiency gains with regulatory expectations around accountability and fairness.
Manulife said the platform it is adopting includes governance and security controls intended to manage how AI agents interact with internal systems. These controls help track how decisions are produced, monitor how data is used, and ensure the systems operate within company policies.
Such safeguards are important in insurance, where automated systems often support processes tied to risk evaluation, claims management, and regulatory reporting.
The operational case for AI agents
For insurers, the appeal of AI agents lies in their ability to reduce manual work across large administrative operations. Claims processing, policy management, internal reporting, and customer support all involve repetitive tasks that often require staff to gather data from multiple sources.
AI systems that can collect and organise information across systems may allow employees to focus more on analysis and decision-making rather than data preparation.
Other financial firms are exploring similar approaches. Banks in the United States and Europe have begun testing AI agents for fraud detection, compliance monitoring, and internal research tasks. In many cases, the goal is not to replace employees but to assist them with time-consuming analysis or data collection.
Research from Accenture’s Banking Technology Vision report suggests that AI-driven automation could help financial institutions reduce operational costs by up to 30% over time, depending on the processes involved. Much of the benefit comes from speeding up routine tasks and improving the accuracy of data handling rather than eliminating large parts of the workforce.
Still, the move from pilots to operational systems carries risks. AI models can produce errors, and automated workflows may amplify those mistakes if they are not monitored closely. That risk is one reason many financial firms are adopting gradual rollout strategies, starting with internal tools before expanding to customer-facing systems.
Manulife’s plan to deploy agent-based AI across its operations shows how large enterprises are testing the next stage of enterprise AI adoption. Rather than focusing on standalone tools, companies are beginning to build systems that can interact with multiple datasets, software platforms, and business processes.
For insurers and banks, the key question will be whether these systems can deliver reliable results while meeting regulatory expectations. If they can, AI agents may become a regular part of financial operations, handling routine work that once required large teams of staff.
As companies push beyond early experiments with generative AI, the focus is shifting to something more practical: making the technology work inside the everyday systems that run large organisations.
(Photo by Joshua)
See also: Agentic AI in finance speeds up operational automation
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
