Meanwhile, a claims adjuster reviews the case not to replicate the AI’s work but to handle the uniquely human elements—customer anxiety, complex liability questions, and edge cases the AI flagged for review. This isn’t just automation; Its a full transferrance from humans to technology to optimise the claims workflow, and excel in the transformation “from risk transaction to customer loyalty”.

The insurance industry, like many other sectors stands between trying to embrace new ideas
and its existing tech estate feeling very outdated and slow to adapt to change. While AI
promises unprecedented efficiency in claims, underwriting, and customer service, many
insurers find themselves caught in what we might call the “automation trap”—using
sophisticated technology to simply speed up traditional processes rather than tring to
challenge the way we use technology to change the way we work The real opportunity lies
not in doing old things faster but entirely new things.

Consider the stark contrast of Haven Life, which now issues life insurance policies in 20
minutes versus the traditional weeks or months, yet customer trust remains a key industry
challenge. Allstate’s AI successfully identifies fraudulent claims patterns, while legitimate
claimants still face frustrating delays. Furthermore, While insurers invest heavily in AI, in fact,
a report from Clearwater Analytics reveals that as much as 74% of insurance companies still
rely on old tech to complete core functions. To make matters worse, up to 70% of insurers’
annual IT budgets are spent on maintaining this outdated technology. However, when they
succeed with AI implementation, the results can be transformative: A Nordic insurance
company automated claims processing with 70% accuracy in document extraction, yet their
biggest challenge wasn’t technical—it was helping employees adapt to managing AI-flagged
exceptions. These examples illustrate how technological capability alone doesn’t guarantee
business transformation.

This transformation is already happening in three fundamental ways, each challenging
traditional insurance paradigms while creating new opportunities for those bold enough to
seize them.

From Annual Policies to Continuous Coverage

Consider Metromile’s AI-driven pay-per-mile insurance model. Rather than annual premiums
based on static risk assessments, their platform monitors driving behaviour and adjusts
coverage accordingly. This isn’t merely a pricing innovation—it’s a shift from periodic risk
assessment to continuous engagement. Zurich UK’s partnership with Sprout.ai, which
achieved 98% accuracy in automated claims processing, demonstrates how real-time data
processing can transform customer interactions from periodic transactions to ongoing
relationships. This shift is already visible across different insurance types. Property insurers
use AI to continuously analyse weather patterns, crime rates, and other dynamic factors to
assess risk in specific areas. For instance, CoreLogic’s AI platforms now streamline
everything from tax verification to flood zone determination, enabling real-time risk
assessment that was impossible under traditional annual review models.
But this shift brings new challenges. When Swiss Re implemented AI for detailed life
insurance risk assessments, they discovered that continuous monitoring required new
technology and skills. Underwriters needed to become data interpreters, combining
traditional risk assessment expertise with the ability to understand and explain AI-driven
insights, again presenting the business with the challenge of having the right skills in the
right place.

From Risk Pool to Risk Prevention

CoreLogic’s use of AI to detect mortgage fraud before it happens points to an even more
fundamental shift: from coverage to prevention. This transformation extends beyond property
insurance. Health insurers use AI to predict potential health issues and guide preventive
care. Auto insurers are using telematics to influence driving behaviour. The insurance
promise evolves from “we’ll pay if something goes wrong” to “we’ll help ensure nothing goes
wrong.” The transformation extends to property insurance, where companies use AI-
powered drones and satellite imagery for proactive risk assessment. This shift to data-driven
prevention fundamentally changes property insurance, though it also raises concerns about
increased policy declines in high-risk areas, although this example presents obvious privacy
issues with lots of image recognition drones flying around.

From Product Provider to Ecosystem Player

The most profound shift is in competitive positioning. Insurance is becoming less about
selling policies and more about orchestrating risk-management ecosystems. MetLife Japan’s
collaboration with Shift AI for fraud detection tripled their success rate not because the AI
was smarter but because it could integrate data from multiple sources in real-time. The
NAIC’s formation of the Third-Party Data and Models Task Force in 2024 acknowledges this
ecosystem reality – insurers increasingly rely on external AI models and data sources. This
creates new challenges around accountability and governance: when an AI-driven decision
involves multiple ecosystem players, who owns the customer relationship? Who bears
responsibility for errors?

This ecosystem play is attracting new competitors. Tech companies with superior AI
capabilities are entering the insurance space, forcing traditional insurers to decide: Should
they become platform players, orchestrating ecosystems of services or specialised providers
within someone else’s ecosystem? Lemonade’s AI claims processor, “Jim,” isn’t just
automation—it’s part of a larger platform play that includes everything from policy selection
to claim resolution. Although evolution always brings new risks. Generative AI models risk
leaking sensitive training data or creating unrealistic expectations. AI-generated “deepfake”
claims challenge fraud detection systems. The NAIC’s guidelines now demand explainability
in underwriting and claims, particularly for third-party AI models.

The Human Element: Beyond Technology

Success in this new landscape requires more than technology investment. When Clearcover
launched its AI customer service platform, it achieved 35% automation within a month. Bu
its real success came from how it positioned AI internally—not as a replacement for human
agents but as a tool to handle routine queries, allowing staff to focus on complex customer
needs. There is a lot of fear fuelled by some of the hype that AI is coming for our jobs, and
we think that the near and medium-term outcome is that it gives skilled people great tools
rather than replacing them.
This points to a crucial truth: AI’s greatest impact isn’t replacing humans but redefining their
roles. Claims adjusters become complexity handlers, underwriters become risk advisors, and
customer service representatives become relationship managers. The challenge isn’t
technological but organisational—how to build cultures that embrace AI as a collaborator
rather than a threat.

Strategic Imperatives for Leaders

For insurance leaders, this transformation demands action on three fronts:

1. Near-term: Start with focused AI implementations in areas like customer service or
claims processing, but design them with ecosystem thinking in mind. To ensure
ethical AI deployment, follow the Geneva Association’s five principles—transparency,
fairness, safety, accountability, and privacy.

2. Medium-term: Invest in data infrastructure and talent development. Success requires
both analytics-ready data and people who can interpret it. Create hybrid roles that
combine domain expertise with AI literacy.

3. Long-term: Rethink your business model. Will you be a platform player or a
specialised provider? How will you balance personalisation with pooled risk? How will
you maintain trust while automating decisions?

Real-world success stories validate this approach. Haven Life’s transformation shows how
balanced implementation works: they didn’t just automate underwriting but reimagined the
entire customer journey, maintaining human oversight for complex cases while using AI to
speed up routine decisions. The examples of MetLife and Clearcover show how balanced
transformation—considering technology, people, and processes together—delivers
sustainable results. Regulatory compliance must be built into AI strategy from the start. The
EU AI Act and NAIC guidelines demand explainability in underwriting and claims. Successful
insurers are turning these requirements into competitive advantages through transparent,
auditable AI systems.

The Path Forward

The future of insurance isn’t about AI replacing humans or merely speeding up existing
processes. It’s about fundamental business model transformation. Success requires
balancing innovation with ethical considerations, efficiency with trust, and automation with
human judgment.
Remember our opening scenario? The real innovation isn’t that AI can process claims
quickly—it’s that the entire system has been reimagined. The claims adjuster isn’t doing less;
they’re doing something different and more valuable. That’s the true promise of AI in
insurance: not just automation but transformation.
The winners in this new landscape won’t be those with the most advanced AI but those who
best understand how to combine technological capability with human insight, ethical
consideration, and strategic foresight. The question isn’t whether to embrace AI—it’s how to
embrace it in a way that transforms your business while maintaining the trust that has always
been at the heart of insurance.

By David Elliman, Global Chief of Software Engineering, Zühlke Group