Artificial intelligence (AI) has been a fixture in the insurance industry for nearly 50 years, used for modeling pricing and managing claim flow. However, the recent evolution, particularly the rise of agentic AI, is ushering in a new era of possibilities, according to Peter McMurtrie, a partner at West Monroe.
“There are really not a lot of players in the insurance space today that are truly fully leveraging AI,” McMurtrie says.
The current trend sees existing automation becoming faster, alongside the emergence of new opportunities focused on transforming the core of the industry. New advancements center on the integration of AI models and capabilities directly into the daily work of underwriters and producers. While the latest solutions are generating significant excitement about the “art of the possible,” McMurtrie says successful deployment hinges on organizations avoiding “AI Theater” and instead implementing fundamental building blocks for successful AI deployment.
The Critical Building Blocks for AI Enablement
For AI to truly take hold, McMurtrie says organizations must address these key structural hurdles.
First, establishing clear governance and guardrails is non-negotiable. This involves defining ethical uses of AI and creating strong policies for data security.
Second, a strong data environment is essential, covering both structured and unstructured data. Since AI is trained on an organization’s data, its availability and granularity are crucial for model effectiveness.
Third, organizations must ensure their technology platform and processes are ready for deployment. This requires moving beyond legacy tech stacks to modern platforms capable of leveraging automation. Furthermore, standardizing workflows and processes is key. McMurtrie cautions against the classic error of automating a bad process, noting that AI makes this challenge even more pronounced, especially when creating custom copilot solutions to align with varied individual workflows.
Cultivating an AI-Ready Workforce
An increasingly critical component in any AI strategy is workforce readiness. As AI interacts more directly with employees, companies must prepare their workforce for adoption and define the future of work.
McMurtrie says preparing a “future-ready workforce” involves two broad strategies:
- AI Native Culture: Exposing the workforce to AI through safe, structured environments, such as internal generative AI solutions, allows employees to experiment, learn, and imagine new ways to leverage the tools.
- Intended Use Clarity: Leaders must ensure teams understand how each tool is meant to be used. Some AI solutions are designed for full task automation, while others serve purely as decision support. For decision support, AI brings relevant information and context – such as a client’s typical coverage needs – but it is not meant to be a prescriptive directive. The expectation is that the professional – be it an underwriter, producer, or claims handler – will apply their own knowledge and experience. Leaders must observe and coach to find the appropriate balance, as ignoring the tool or following it blindly both result in poor outcomes.
Avoiding “AI Theater” and Focusing on Under-Realized Use Cases
While many firms engage in “AI Theater” – in which executives spend excessive time talking about moonshot efforts – the highest-value opportunities are currently under-realized.
“Where you really have a lot of power coming in is around workflow orchestration and interconnecting the AI models and capabilities into the workflow of the underwriter or the producer,” McMurtrie explains.
McMurtrie says it is crucial that firms prioritize high-value AI use cases – especially AI copilots for complex underwriting, complex claim handling, and producer selling – that bring just-in-time unstructured data and context to decision makers
This approach leads to better decisions, delivered faster, without the need to sift through multiple systems to search for information. McMurtrie believes that Improving decision velocity and quality for the organization’s highest-paid associates represents the highest potential value use case.
Spotlight on Governance
The governance landscape is marked by a notable split: internal caution is slowing adoption, while external regulation has lagged. McMurtrie fully expects to see an increase in regulatory guidance and control from bodies like the NAIC emerging in the coming months and by year-end.
Internally, governance controls currently act as a “governor,” impeding fast movement due to concerns over disparate treatment and decisioning, as well as data security. The recent rash of cyber attacks and data exfiltration within the insurance industry has only heightened this caution around exposing data to models. Organizations must accelerate their internal controls, but the incoming regulatory controls are expected to create a “challenge state” that requires cautious adoption for at least the next 18 months.
Distinguishing Success: ROI and Discipline
Currently, many organizations report a low return on investment (ROI) for their AI initiatives. McMurtrie says this low ROI is directly tied to the failure to implement the necessary building blocks. Specifically, ROI is low where there is no clear business case, poor underlying data quality, inconsistent workflows, or a lack of intentional change management and people readiness.
Conversely, where these elements are addressed, McMurtrie notes the return is exceptionally high – up to 13 multiples of the average ROI. Differentiation is rooted in discipline: successful firms prioritize solving tangible business problems, framing AI as a tool that fits into a broader business strategy, rather than seeking an “AI strategy” or a “moonshot”.
McMurtrie says the top-quartile performing property and casualty carriers have consistently made disciplined investments in their technology, data and operations. This structure allows them to be targeted and consistent in their AI adoption and differentiates them from the laggards. Ultimately, McMurtrie believes firms that match ambitious goals with targeted and methodical organizational processes will see the most value from their AI initiatives.