Just as we were coming to terms with Artificial Intelligence (AI), the next step forward changed the game. Agentic AI offers robots that can simulate humans by interacting with each other and other applications. Could this replace much of the actuarial work? Will we see Actuarial Modelling Agents and Financial Reporting Agents? Our fifth article looks further into the opportunities.
Agentic artificial intelligence is an important advancement in the field of AI. Unlike traditional large language models (LLMs), agentic AI systems can do more than perform language-based tasks such as writing or coding; they are also able to make decisions and to utilise tools to interact with their environments. These systems can pursue defined goals, adjust to changing situations and incorporate feedback. This represents a significant step towards developing AI that more closely resembles human decision-making and autonomy.
In effect, these systems display a degree of initiative. You can think of them as virtual workers, each designed to fulfil different roles. Rather than being limited to scripted interactions such as chatbots or following the step-by-step instructions of automation tools, they are able to pursue broader goals and adapt the actions they take to changing circumstances.
What makes AI agents truly transformative is their ability to work together within multi-agent systems. Relying on a single, all-encompassing agent to complete a complex task can make that agent overly complicated and difficult to manage. A more effective approach is to design a multi-agent system, where a central project manager agent coordinates the overall process and delegates specific tasks to specialised sub-agents. Each sub-agent focuses on a particular aspect of the problem, in the same way as members of a human team work together to achieve a common goal.
With the ability to deal with complex tasks in real-time, and to use tools to achieve goals, AI agents have clear potential to
This is a very new space and many organizations are still exploring the potential. We expect to see AI agents deployed in a number of areas, including:
Clearly, they could go further, including setting actuarial assumptions, defining stress scenarios or even supporting the full end-to-end financial reporting cycle. However, most regulators and professional bodies have been clear that AI outputs require human oversight and validation This will likely mean they first appear in more of a monitoring role and suggesting improvements; AI agents are some way off contributing to the critical path of financial reporting.
There is also considerable interest in whether agentic AI could produce whole actuarial models from scratch for life businesses. While there is no theoretical reason why not in the medium term, concerns around transparency, auditability and validation mean that widespread adoption of such technology is unlikely anytime soon. Nonetheless this remains an area to watch as the intelligence and accuracy of the models improves day by day.
This summary was correct as of December 2025.
We actively monitor developments in technology and research how they may be applied to financial modelling in a sustainable way. Talk to us to find out more about how this technology can help your business.
Email software.enquiries@wtwco.com and we’ll connect you to our local experts.
