Agentic AI is a step beyond traditional automation. These systems can act with a degree of independence: carrying out sequences of tasks, adapting to feedback and pursuing defined goals.
In insurance, that means moving from technology that simply supports processes to tools that can proactively gather data, triage cases or trigger workflows. Work that once absorbed hours of underwriters, claims handlers or operations teams is now completed faster, more consistently and at scale.
The technology is still new. Early pilots began in 2024, typically in controlled environments such as underwriting support or claims triage. By early 2025, only a minority of insurers, around 10-15% in the London market and closer to 20% across EMEA, had started experimenting. Adoption so far is measured in months, not years, and most firms remain in pilot mode rather than operating at scale.
Why Does Agentic AI Matter for Insurance?
Insurance is built on risk, trust and regulation, three things not easily handed to machines. But agentic AI differs from older systems that simply generate outputs. These agents act, decide and follow through on goals. This is a fundamental shift, moving AI from a back-office tool to an operational partner that supports underwriting, claims and workflows.
The question is no longer whether this shift will affect insurance, but how far and how fast. For firms under pressure to do more with less, the attraction is clear: faster decisions, fewer errors and better use of human time. Agentic AI will not replace judgment or empathy, but it will give people the space to focus where those qualities matter most.
Early impacts on the insurance market
We are beginning to see genuine momentum across the market, even if progress is still at an early stage. In April 2025, the Lloyd’s Market Association (LMA) published results from a market-wide survey showing that 14% of London market firms have already deployed or tested agentic or generative AI in underwriting. This indicates that the shift from theory to practice is underway, particularly in high-value and decision-heavy areas.
Elsewhere, some syndicates are trialling agent-led triage and data gathering to support underwriting. These tools are not yet replacing underwriters but helping them to move faster, reduce manual follow-up and devote more time to assessing risk.
Across the wider EMEA region, McKinsey reports that around 20% of insurers are piloting or planning to pilot agentic AI within the next year, with activity concentrated in claims, underwriting and finance.
The technology is arriving, and the appetite to test it is growing. What is emerging is a clear pattern - insurers are prepared to experiment in high-value, decision-heavy processes where speed and accuracy make the greatest impact. The question is whether those early pilots can scale into consistent business value.
What are the benefits of working with Agentic AI?
Where agentic AI is embedded effectively, the benefits are already visible. Firms report cost efficiencies, faster product cycles and stronger broker engagement. In commercial lines, brokers describe quicker responses and faster quote turnaround, not because underwriters are working longer hours, but because agents are removing operational friction.
The numbers back this up. McKinsey suggests that insurers using agent-led underwriting support can cut processing times by 30-40%, while EY highlights that autonomous claims handling can reduce average handling costs by up to 20%. Broader studies show task times reduced by almost 70%, freeing teams to spend more of their effort on higher-value work, alongside a lift in job satisfaction.
Customers are beginning to see the difference, too. Routine queries, policy changes and straightforward claims are increasingly handled by autonomous agents, easing pressure on service teams and accelerating resolution.
For early adopters, this is more than a technology trial. It sends a clear market signal that they are digitally fluent and operationally agile. Munich Re has piloted agent-led data extraction for underwriting submissions, QBE has explored autonomous agents in claims triage, and Convex has openly invested in generative and agentic AI as part of its digital-first strategy.
As one industry leader put it:
“The real value is not in replacing underwriters or claims teams, but in giving them the time and clarity to focus on the risks and relationships that matter.”
What are the drawbacks of Agentic AI in insurance?
Progress is not without challenges. Three issues are raised most often: governance, speed and organisational readiness.
Governance remains paramount. Any AI system operating in a regulated market must be explainable, auditable and compliant. For underwriters and claims teams, whose value rests on professional judgement, handing decisions to autonomous agents raises real questions of accountability.
The second barrier is speed. Technology is advancing so quickly that by the time a use case has been scoped, tested and approved, the underlying platforms may already have shifted. This creates tension between innovation teams keen to push forward and risk functions focused on stability and compliance.
The third is organisational. Moving from linear processes to agent-led workflows requires more than new technology. It demands new governance structures, fresh ways of working and often, a shift in culture.
Layered on top are practical hurdles - a shortage of in-house expertise, uncertainty over which vendors can be trusted, and the need to ensure tools are genuinely fit for the insurance market. Taken together, it is clear why some firms are cautious. This is not simply a technology upgrade, but a transformation in how insurance work is organised and delivered.
Athula Alwis – CEO, AllDigital Specialty Insurance comments:
“We’re not talking about replacing underwriters. We’re talking about enhancing selection, scaling decision‑making, and bringing speed to something that used to take days.”
Source: Insurance Business
What does this mean for executive teams?
Leadership is critical. This is not just a CIO challenge or a line in the digital strategy, but a fundamental business shift. The CEO must set the tone, ensuring innovation is grounded in trust and transparency. The COO and CIO need to determine how to embed autonomous agents safely into operations without disrupting established processes. CUOs and Claims Directors face the task of modernising their functions without losing the judgment and expertise that make them valuable.
Boards and executive teams also need to establish clear ethical and governance frameworks. If an autonomous agent makes a decision that affects a customer, who owns the outcome? This is no longer a theoretical question. It demands a clear answer.
Which AI tools are in use in insurance?
The tools are improving fast. Open-source frameworks like LangChain and AutoGen are already being used for prototyping agent-based workflows. On the enterprise side, Microsoft Copilot Studio, Cognosys and Adept AI are pushing compliance-aware solutions into the market. In a notable move, ServiceNow announced its $2.85 billion acquisition of Moveworks in March 2025, a clear signal of intent to accelerate enterprise-scale Agentic AI deployment across functions such as IT, HR, finance and customer operations.
Closer to home, platforms like Guidewire and Shift Technology are starting to integrate agentic features into their insurance suites. And several startups are building point solutions for tasks including onboarding, triage and reporting.
It’s a crowded space, and not everything will stick. Due diligence is non-negotiable. Any tooling needs to align not just with your stack, but with your governance, risk appetite and customer promise.
The future of AI in insurance
Agentic AI is not simply a smarter form of automation. It represents a step change that could reshape how the insurance business operates, from process to people to performance. The firms that move early, experiment carefully and build strong foundations will not only stay ahead of the curve, they will help define the next chapter of the industry.
About the Author
Meera Joshi heads the Insurance Practice at Freshminds, partnering with boards and executive teams to steer transformation across the sector. With more than a decade in recruitment and consulting, she has a track record of helping organisations secure the leadership and strategic talent they need at moments of change.
Her work sits at the intersection of people and technology, supporting insurers as they respond to shifts such as the rise of agentic AI. Her perspective is grounded in years of collaboration with insurers, brokers and market associations, enabling them to build teams that deliver today while preparing for the challenges of tomorrow.
