Why insurance innovation keeps failing — and how mea Platform is changing this

Despite years of digital transformation and rising AI adoption, insurance operations remain largely unchanged, but mea Platform believes it is changing the game.

Across the global insurance value chain, operational spend still accounts for an estimated 12–14 cents of every premium dollar, while only 3–5 cents is typically directed toward activities that generate true competitive differentiation. Despite billions invested in digital transformation programmes and, more recently, artificial intelligence, much of the industry continues to rely on fragmented systems and manual workarounds. Additionally, many of the highest paid (and most important) talent in the industry still spend the majority of their time on manual workflows.

Speaking to FinTech Global, as part of the prestigious AIFinTech100 list, mea Platform’s Chief Marketing Officer Elliott Bundy argued that he sees the issue as structural rather than technological.

“The insurance industry made a fundamental bet that spending on processes and the back office would create differentiated performance. It hasn’t. The bet was wrong, and today, competing on the back office means losing the opportunity to work where it matters,” says Bundy.

Instead of rethinking how insurance work is executed, most firms have digitised existing processes, layering new tools on top of legacy infrastructure without fundamentally changing how decisions are made or workflows are run.

That gap between ambition and execution is where a new category of insurance-native platforms is beginning to emerge in the form of operational infrastructure, rather than the more traditional and more loosely defined “tools”.

What’s different now

The insurance industry, traditionally slow to evolve, has embraced AI faster than almost any previous wave of technology. But adoption has not yet translated into performance transformation.

Across carriers, brokers and MGAs, AI programmes tend to follow a familiar pattern: rapid experimentation, tightly defined use cases, and limited progression into production at scale. On paper, the activity signals momentum. In practice, very little changes in how the industry actually operates.

“There’s huge pressure to get AI programs up and running,” says Bundy. “Pilots and quick wins are fantastic. But the chance to create solved problems is now bigger than it ever has been — the broader opportunity is to work differently.”

And that distinction matters more than it first appears.

Point solutions such as document processing, triage and data extraction can make a big difference in efficiency gains. But they might sit inside operational departments that are working in silos. In the insurance value chain broking links to underwriting, which links to claims, which links to finance, and each function still depends on fragmented layers of technology that do not fully connect.

“AI often becomes an enhancement layer rather than an operational reset,” Bundy says. So even as pilots succeed, the broader structure holds.

And so, the pattern repeats. Projects demonstrate technical viability, but struggle to scale beyond isolated pockets of the business.

What gets missed, in his view, is the nature of the opportunity itself.

“Chances are you are missing the bigger opportunity,” he says, “of not just embedding AI into existing solutions, but rethinking how the workflow itself should operate.”

Because once AI is introduced into a system built for human execution, it inherits the constraints of that system. It speeds things up, but it does not change what the system is.

Bundy argues that this is the core mistake, and one that mea Platform can solve. Rather than layering intelligence onto existing workflows, organisations can invert the model entirely, designing workflows around automation from the outset.

In that structure, AI is no longer an assistive layer sitting beside human work. It becomes the execution layer itself, handling repeatable operational tasks end to end.

“The AI takes care of the repetitive, mundane tasks, and the human workforce takes care of the consequential,” Bundy says.

And that minor tweak in language is incredibly significant in implication, and where real change is beginning. Because the question is no longer whether AI works inside insurance. It is whether insurance is willing to be rebuilt around how AI works.

Building purposefully for insurance

If insurance is to be rebuilt around how AI works, then the technology itself needs to understand the industry it is being asked to transform.

That belief sits at the heart of mea Platform’s approach. While much of the market has focused on applying increasingly sophisticated AI models to existing insurance processes, the company took a different path. Rather than adapting generic technology to the sector, it set out to build an AI-native platform designed specifically for the language, workflows and operational realities of insurance itself.

For Bundy, that distinction is fundamental.

“We really try and understand the problem our clients are trying to solve,” he explains.

“From there, we then go and apply technology to those problems, versus being a technology-first shop that is going in search of problems to solve.”

That philosophy shaped the platform’s development from the outset.

The company spent years building an insurance-specific language model and knowledge graph designed around the industry’s terminology, processes and decision structures. The goal was not to adapt a general-purpose AI model to insurance workflows, but to create technology purpose-built for the sector.

The rationale was straightforward. Insurance is filled with nuance. A policy placement, a claims submission and a bordereaux reconciliation may all sit within the same value chain, but each carries its own language, context and operational complexity.

That decision now sits at the centre of the company’s proposition, and partners are waxing lyrical at the benefits.

“The choice to partner with mea means you’re getting a solution that is pre-trained in the language and workflow of insurance operations,” Bundy explains. “As a result, it is very quick to integrate, and the ROI is fast to achieve.”

The results are already being demonstrated at scale. mea Platform has more than 30 live deployments across 20+ countries and has processed over $400bn in gross written premium through its platform. Customers have reported a 30% increase in broker productivity, a 40% average increase in underwriting capacity and reductions in operating costs of up to 60%, helping insurers redirect resources towards activities that drive competitive differentiation.

For insurers, the appeal is the prospect of avoiding the lengthy integration cycles that have accompanied so many previous technology deployments. And that is where mea sees its greatest distinction.

“It has that ability to just work out of the box,” Bundy says. “That’s a huge differentiator for us. It’s something that’s very hard to replicate from many competitors.”

The ambition, however, extends beyond efficiency gains.

What mea has ultimately built is a system capable of understanding insurance in its own language, operating within its own logic, and executing work across multiple stages of the value chain.

Eyeing new opportunities

For mea Platform, the opportunity ahead extends far beyond automating individual processes or improving operational efficiency. The company’s vision is centred on helping insurers shift their focus away from the mechanics of running the business and towards the activities that create genuine competitive advantage.

“We really think there’s an opportunity to rationalise that spend that has traditionally gone into operations or the back office,” says Bundy. “We think of this as the opportunity to finally say back office is a solved problem.”

It is an ambitious proposition, but one that reflects a broader shift taking place across the industry. As AI capabilities continue to mature, the conversation is beginning to move beyond experimentation and towards execution at scale.

For mea Platform, that means continuing to develop an operating layer capable of supporting the full lifecycle of insurance operations: across underwriting, claims, finance, and broking, not as a point solution bolted onto one part of the workflow. The ultimate goal is not to remove human expertise from the process, but to elevate it.

If the last decade was defined by digitisation, the next may be defined by execution. And for mea Platform, that represents the beginning of a new era for insurance operations.

“From the point of view of a mea customer, we can allocate our most valuable time, and our people’s most valuable time, to creating differentiated performance.”

This article was written by the editorial team at FinTech Global.