AI can help close re/insurance blind spots; sharpen growth, risk and returns – mea

As re/insurers explore AI’s potential beyond automation, Fernando Villar of mea Platform challenges the industry to rethink the opportunities it is missing and the operating model holding it back.

Key points:
Operational gaps can cost good business
AI can support decisions, not just speed
Earlier insight can change underwriting outcomes

By Fernando Villar, managing director, mea Platform

Imagine a reinsurer losing a promising opportunity without ever deciding to turn it down. A broker has sent the risk, it appears to fit the appetite and the expertise to assess it is there. But information needs assembling and the underwriting team is stretched. By the time the file is ready, the opportunity is gone. Nobody made the decision.

I have been working with brokers, insurers and reinsurers, discussing their strategies and operating models. The challenge is to grow profitably without allowing costs and complexity to rise at the same pace. Does the operating model enable the strategy, or quietly limit it?

Recent ISG research gives that question weight. Carriers estimated that roughly one in nine broker submissions is declined or left unquoted because of operational constraints, despite an appetite for the risks. 

Missed opportunities

A deliberate “no” can protect more value than an enthusiastic “yes”. Greater capacity should expand underwriting choices, not lower standards. We cannot control the market cycle, but we can improve our response.

Even declined business offers intelligence. “Lost on price” should prompt questions: were our terms unsuitable or our response too slow? Repeated requests for protection we do not offer could suggest a new proposition. This is the unwritten portfolio: learning from opportunities declined, lost or never developed. Its value rests on our records and permitted feedback, not AI-generated guesses.

That need to learn does not end when we accept the business. Imagine a casualty treaty approaching renewal. The latest claims report shows reserve increases on several older claims. If the report is checked against previous figures and underwriting assumptions only after renewal, the information arrived in time, but the understanding did not.

Earlier insight

Earlier analysis could have changed pricing, terms or participation or confirmed that the proposed renewal was sound. Technology cannot remove the uncertainty of claims development. But information already received should reach decision-makers while there is still time to act.

Both blind spots narrow management’s choices. Addressing them can strengthen risk selection, capital allocation and client relationships, as well as reduce expenses. This is about where and how we compete, not simply how we process work.

Agentic AI can help carry that work forward. A controlled workflow could compare claims movements, check treaty references and bring material changes, with source evidence, to the responsible specialist. The same approach can connect submission intake, appetite checks, underwriting referrals and financial reconciliation. The test is whether the next person can act, not simply whether the previous task ran faster.

Agents need rigorous testing, clear permissions, traceable evidence and recorded actions, with referral when confidence or authority is insufficient. Consequential decisions must remain with accountable people.

Common data standards and controls should support a consistent view of risk while respecting local requirements. Local teams must help shape implementation.

At mea, we bring more than agentic solutions. Our team combines technology expertise with practical experience of re/insurance operations. That matters when clients need partners who understand their business, work with existing systems and deliver measurable value. For a defined first workflow, I would seek that value within a quarter, with realistic milestones and predictable costs.

The strategic choice is not simply what to buy or build, but where to focus scarce expertise and capital. Proven technology can take on repeatable work, allowing teams to concentrate on what differentiates the business. Yet technology alone does not deliver results. It requires clear accountability, integration into daily operations and people ready to work differently. A successful pilot demonstrates potential. The real test is whether it delivers better decisions, stronger client relationships and profitable growth.

A deliberate ‘no’ can protect more value than an enthusiastic ‘yes’.

From pilots to scale

As 2027 approaches, I would scale a few high-impact workflows rather than multiply pilots. Measure success through opportunities assessed sooner, earlier portfolio insights, strong controls and lower costs. Ultimately, the prize is disciplined underwriting, profitable growth and stronger risk-adjusted returns.

At Baden-Baden a question deserves a place at the table: what opportunities never reach our underwriters, and what do we discover about our portfolios only after the moment to act has passed?

The real opportunity is to rethink how we operate, using agentic AI to connect information, expertise and decisions across the reinsurance value chain. The aim is to help brokers place risks more effectively, reinsurers make better decisions and clients benefit from faster, more tailored solutions and better outcomes.

Fernando Villar is a managing director at mea Platform and can be reached at: fernando.villar@meaplatform.com

For more news from Baden-Baden Today, click here.

Did you get value from this story? Sign up to our free daily newsletters and get stories like this sent straight to your inbox.