Insurers are moving quickly to adopt generative AI in the form of chatbots, summarizers, document analyzers, and recommendation tools across underwriting, claims, and servicing.

While these innovations deliver immediate value by automating tasks and improving productivity, most are only isolated features instead of a part of a cohesive intelligence strategy. Insurers risk creating a patchwork of smart tools that don't learn from one another or scale strategically. They just…exist, with the whole never becoming greater than the sum of its parts.

The next wave of competitive advantage won't come from adding more gen AI features. It will depend on a shared, intelligent foundation that spans all core systems and learns across the enterprise. In fact, the Geneva Association now encourages insurers to invest in strong data infrastructure and hybrid architectures if they hope to execute productive AI deployments.

Isolated solutions won't cut it. To unlock sustainable business value, insurers must construct gen AI with broad foundations.

Gen AI Features vs. Foundation

Gen AI features – each one built to solve a narrow issue – are all the rage. These tools, limited in scope, include claims processing tools, underwriting analysis, suggestion engines, and, most notably, customer service chatbots. Indeed, conversational AI chatbots are already widely used to handle routine policyholder interactions, offering real‑time assistance through intelligent chat and voice interfaces.

Yes, these tools are valuable. They're also disconnected.

These systems can execute tasks but cannot communicate with each other – not sharing insights, not learning collectively, not evolving as an "AI suite" greater than the sum of its parts. A gen AI foundation, by contrast, provides a shared intelligence layer that continuously learns from every interaction and use case. This transforms isolated automation into collective intelligence, where each new AI capability strengthens the whole.

The Cross-System Reality: No Hyper-Personalization

Most insurers operate in complex, multi-system environments, guided by multiple policy administration systems (PAS) with decentralized claims, billing, and CRM platforms. Each system contains only a partial perspective, creating dangerous silos across the policy lifecycle that don't empower insurers to offer hyper-personalization or satisfying customer journeys.

A true gen AI foundation does not replace these systems of record. It connects and contextualizes them through a unified intelligence layer. In fact, recent trend analyses emphasize that a unified semantic layer is essential for contextualizing disparate insurance data sources and enabling consistent AI reasoning. The objective is not just data access – it's shared understanding that evolves across systems, functions, and time.

The Brain of Record: Systems to Synapses

Traditional systems of record (SoR) are designed to manage both static data and dynamic data, including real‑time transaction details, plus historical data, to support audits and compliance checks.

Alternatively, a brain of record (as insurance innovation leaders call it) goes a step further by capturing understanding, not just information – integrating structured and unstructured data, maintaining lineage and traceability, and enhancing data with learned relationships and insights.

This living intelligence and memory is constantly evolving, able to categorize insights, learn from interactions, and organize context from individual transactions to enterprise-wide patterns. Metadata plays a foundational role in this foundation layer by tagging every piece of information with its source, purpose, and relationship.

AI can reason using that context and generate its own metadata, identifying new concepts, clusters, and emerging connections across the enterprise. This helps create a seamlessness appreciated by employees and customers.

How to Unify Two Worlds

Application programming interfaces (APIs) provide access, but not context.

These platforms offer only a narrow transactional view of policies, claims, or payments, but cannot connect insights across systems or reason over time. Similarly, traditional data warehouses – optimized for storing and querying structured business data, but not for handling high‑scale, real‑time, AI‑driven workloads – can centralize information, but can't execute intelligence or reasoning.

An overarching AI data foundation unifies both worlds by continuously synchronizing structured and unstructured data. It also expands the function of metadata by identifying new risk categories, linking similar claims, and surfacing emerging relationships. This foundation enables real world scenarios, such as AI-driven underwriting agents that proactively assess renewals, detect emerging risks, and support human decision-making with contextual intelligence.

But building this kind of gen AI foundation is not a plug-in exercise. You can't fake it. It requires robust AI data and context fabric that is dynamic, hybrid, layered, governed and evolving. This evolving fabric should act as the backbone of enterprise intelligence.

Why the Foundation Matters

Insurers are moving from siloed deployments to enterprise‑wide AI platforms, where generative and multimodal models simultaneously improve claims, underwriting, and customer engagement by sharing insights across functions. This shift provides more consistent, real‑time intelligence across the value chain by creating a shared context across underwriting, claims, and servicing departments.

AI data foundations also empower insurers with three huge benefits:

1. Adaptive intelligence: AI systems that get smarter with every interaction
2. Governed trust: every insight is explainable, traceable, and compliant
3. Scalable reuse: each new AI use case strengthens the shared foundation rather than creating another silo

This effectively turns AI from a collection of decentralized tools into an engine of end‑to‑end enterprise intelligence.

Ride the wAIve

The next wave of AI in insurance will be created by the collective depth of organizations' data foundations, not the number of features they host.

Insurers that invest in a brain of record – an evolving intelligence layer that learns, organizes, and grows alongside the business – will unlock lasting competitive advantage. Though complex, this data architecture is very achievable with the right integration strategy and governance expertise.

The insurers that build a strong foundation for their AI today will define the intelligence standards of the industry tomorrow.

Artificial intelligence (AI) is no longer an experimental technology in the insurance sector. By 2026, AI in insurance has become a core driver of underwriting accuracy, claims automation, fraud detection, and customer personalization. Insurers worldwide are leveraging artificial intelligence in insurance, predictive analytics, and machine learning to transform traditional operating models into intelligent digital ecosystems.

However, while AI delivers measurable benefits, it also introduces risks and ethical considerations that insurers must manage carefully.

1. The Impact of AI in Insurance 2026

The impact of AI in insurance extends across the entire value chain — from policy issuance to claims settlement.

In 2026, insurers are using AI-powered systems to:

• Analyze real-time risk data
• Automate underwriting decisions
• Accelerate claims processing
• Detect fraud patterns instantly
• Personalize insurance products

AI-driven platforms process vast amounts of structured and unstructured data in seconds, enabling insurers to make faster and more informed decisions. The shift from reactive to predictive operations has significantly improved operational efficiency and customer satisfaction.

2. Advantages of AI in Insurance

The adoption of AI insurance solutions brings multiple strategic advantages.

Improved Underwriting Accuracy

AI in underwriting uses predictive analytics to evaluate risk factors with greater precision. Machine learning models analyze historical claims, behavioral data, IoT inputs, and demographic insights to generate accurate pricing models.

Faster Claims Automation

AI claims automation reduces manual review processes. Image recognition, natural language processing (NLP), and intelligent workflows allow insurers to approve simple claims within minutes.

Fraud Detection Enhancement

Fraud remains a major challenge in insurance. AI-powered fraud detection systems identify anomalies, suspicious behavior patterns, and claim inconsistencies more effectively than traditional rule-based models.

Cost Reduction and Efficiency

Automation minimizes administrative overhead, reduces processing errors, and lowers operational costs.

Personalized Customer Experience

AI enables insurers to offer tailored coverage recommendations, proactive risk alerts, and 24/7 chatbot support, improving customer engagement and retention.

These advantages position AI as a competitive differentiator in 2026.

3. Effects of AI on Insurance Operations

The operational effects of AI in insurance are transformative. Core processes such as underwriting, policy servicing, billing, and claims are becoming increasingly automated and data-driven.

Insurers are moving toward:

• Cloud-based AI platforms
• Real-time risk modeling
• Embedded insurance powered by APIs
• Data-driven decision frameworks

AI integration also supports better regulatory reporting and compliance management through automated monitoring systems.

As a result, insurers can launch new products faster, respond to market changes more effectively, and maintain stronger operational resilience.

4. Disadvantages and Challenges of AI in Insurance

Despite its benefits, AI adoption in insurance presents challenges that cannot be ignored.

Data Privacy and Security Risks

AI systems rely heavily on customer data. Ensuring compliance with global data protection regulations is critical.

Algorithm Bias

If AI models are trained on biased data, they may produce unfair or discriminatory outcomes in underwriting or claims decisions.

High Implementation Costs

Developing and integrating AI insurance platforms requires significant investment in infrastructure, technology, and skilled talent.

Workforce Disruption

Automation may reduce certain job roles, requiring workforce reskilling and organizational change management.

Regulatory and Ethical Concerns

Regulators increasingly scrutinize AI decision-making processes to ensure transparency and accountability.

To mitigate these risks, insurers must adopt responsible AI frameworks, robust governance models, and continuous monitoring systems.

5. The Future of AI Insurance Beyond 2026

Looking ahead, AI in insurance will become even more sophisticated. Emerging trends include:

• Real-time underwriting using IoT and telematics
• Advanced climate risk modeling
• AI-powered conversational insurance platforms
• Blockchain integration for secure and transparent claims
• Autonomous risk assessment engines

Machine learning in insurance will continue evolving, enabling smarter risk pricing, improved loss prevention strategies, and enhanced customer-centric solutions.

Insurers that strategically invest in AI-driven digital transformation while maintaining ethical standards will lead the industry.

Conclusion

AI in insurance 2026 represents both opportunity and responsibility. The impact of artificial intelligence on underwriting, claims automation, fraud detection, and customer engagement is undeniable. Advantages such as efficiency, personalization, and predictive risk management are driving widespread adoption.

However, insurers must carefully manage disadvantages including data privacy concerns, algorithm bias, regulatory complexity, and workforce disruption.

In the coming years, success in AI insurance will depend not only on technological innovation but also on governance, transparency, and trust. Insurers that balance innovation with responsibility will shape the future of the industry.

While the excess and surplus lines market was once an option of last resort, today it is all too frequently a first step in the process of insuring risk.

A coastal property facing new catastrophe models. A business navigating cyber exposure. A specialized liability account that has outgrown admitted guidelines. For agents and brokers, E&S is now part of everyday operations.

This shift has forced the distribution side of the industry to move faster, communicate more clearly and operate with greater precision. Technology is becoming a bridge to help insurance agencies keep pace while also strengthening relationships with their carrier partners.

Speed matters, but clarity matters more

Unlike admitted markets, where rates and underwriting changes can take time to filter through regulatory processes, non-admitted appetites can shift quickly based on loss trends, capacity and real-time market conditions. A carrier writing a class of business today may pull back tomorrow or adjust pricing as results demand it.

For the retail agent, this reality creates a constant challenge: Where does this risk belong currently? In the past, the answer has required multiple submissions, follow-up emails and trial-and-error market shopping. That cycle slows service, strains staff resources and frustrates underwriters who receive incomplete or misrouted submissions.

Technology can help avoid this cycle of frustration and wasted time, not by replacing relationships but by reducing the friction that can damage them.

Streamlined placements support better partnerships

Modern E&S placement platforms are designed to make submissions cleaner, faster and more consistent. The best tools help agents submit once, validate completeness and route risks to the right markets based on current appetite.

This kind of upfront triage benefits all involved. Agents spend less time chasing dead ends. Underwriters spend less time sorting through half-built submissions. Carriers receive applications closer to their appetites, with clearer exposure data and fewer missing pieces.

The result is a more efficient exchange that respects the time and expertise on both sides of the relationship. We’re seeing that with the deployment of Xchange - Powered by SIAA, which provides our members a faster, cleaner and easier way to access and place E&S business. 

Reducing errors and improving underwriting confidence

One of the most persistent challenges in E&S is submission accuracy. When clients want fast answers, agency teams sometimes make assumptions to move the process along. These seemingly educated guesses can create big delays later when an underwriter must circle back for corrections.

Technology that enriches submissions with third-party data sources can reduce the burden. Property records, hazard data and other verification tools can help confirm details before the submission ever reaches the carrier.

Doing this leads to fewer surprises, fewer resubmissions and a smoother path to a quote. More importantly, it helps carriers trust what they are seeing, which ultimately contributes to stronger carrier-agent relationships.

AI should be an optimizer, not a replacer

Artificial intelligence is playing a growing role in the E&S workflow, but the industry must be clear-eyed about its realities.

AI can help organize information, identify inconsistencies and accelerate routing. It can reduce manual data entry and make it easier for agents to package risks in an underwriter-ready format.

What it cannot do is replace underwriting judgment.

Complex accounts still require human experience, context and expertise. Technology works best when it clears away administrative clutter so underwriters and agents can focus on conversations that matter: coverage structure, risk controls, exclusions and long-term strategy. When positioned correctly, AI supports relationships rather than threatening them.

Strengthening carrier relationships through better submissions

Carrier relationships are built on trust, consistency and professionalism. In the E&S space, where underwriters face heavy submission volume, standout agencies are those that deliver clear narratives and decision-ready accounts. Technology helps agencies meet that standard at scale.

By standardizing intake, improving exposure clarity and managing workflow discipline, agents become better partners to their markets. Carriers benefit from lower acquisition expense per policy, improved risk selection and fewer wasted cycles.

Over time, these operational advantages translate into stronger long-term collaboration.

Carriers tend to prefer distribution partners who can deliver reliable data quality and efficient servicing without requiring carriers to expand headcount at the same rate as submissions.

Agencies that adapt will protect their growth

For agents and brokers, the risk of ignoring technology is not about missing a trend. It is about falling behind both the market and the competition.

As risks become more complex, turnaround time is becoming a competitive differentiator. Agencies relying solely on inbox-driven workflows will find it harder to shift books of business, maintain service levels and compete for talent.

The goal is not to adopt technology for the sake of shiny tech solutions. Rather, the goal is to protect the value of the agency by making the placement process faster, cleaner and easier to hand off to the next generation.

Relationships remain at the center

E&S will always involve more complexity than standard business. But complexity does not have to mean inefficiency. With the right technology, agents and brokers can keep pace with a rapidly evolving market while building better carrier relationships through stronger submissions, smarter routing and clearer communication.

The future of E&S distribution will not be defined by replacing people. It will be defined by empowering them. When technology reduces friction, relationships have room to grow.

We've all had a chuckle about the occasional hallucination by generative AI: the time when it recommended using glue to keep cheese from sliding off a piece of pizza; when an Air Canada chatbot promised a passenger a bereavement fare despite a policy to the contrary, and the airline had to live up to that promise; when a lawyer unknowingly submitted a brief to a judge that was based on citations from court cases that never happened; and so on. 

But a couple of recent stories go well beyond the chuckle level. While generative AI continues to show all the promise in the world, these stories demonstrate consistent problems that, unchecked, would lead to severe consequences. 

Let's start with the one about war games in which large language models almost always recommended escalating to nuclear weapons.

As Axios reports, a researcher at Kings College London pitted three popular LLMs — GPT-5.2, Claude Sonnet 4 and Gemini 3 Flash — against each other in 21 war games in which the AIs acted as the leaders of major nations. The scenarios included threats to survival, but also included lower-stakes conflicts, such as border skirmishes and resource competition. Yet 95% of the time, at least one of the LLMs "used" nuclear weapons, and escalation typically ensued. 

For anyone without a strong Dr. Strangelove streak, those results reflect a scary misjudgment. While the U.S. and the Soviet Union considered tactical nuclear weapons to be legitimate parts of their arsenals in the early years of the nuclear age, those were also the times when the countries casually considered using nuclear weapons for industrial uses such as mining and natural gas extraction. It's been clear for decades that nuclear weapons are simply too powerful for their effects to be limited to legitimate military or industrial targets. 

Even at one kiloton, the smallest payload for what's considered a tactical nuclear weapon, the explosion would be 100 times as powerful as the biggest conventional bomb in the U.S. arsenal. At the top end of the range for a tactical nuclear weapon (generally considered to be 100 kilotons), the explosion would be some seven times as powerful as the bomb dropped on Hiroshima, which destroyed a military target but also killed an estimated 140,000 people, the vast majority of them civilians. The radiation released can also reach far beyond the targeted area. 

While the Kings College researcher noted that no one is handing AIs the keys to nuclear weapons systems, he said, "Militaries are already using AI for decision support — and research suggests those systems may lean into rapid escalation under pressure."

The other article that caught my eye relates to ChatGPT Health. The app, launched in January, is consulted by some 40 million people every day — and a study found the potential for major problems with the app's diagnoses. For more than half of the study's hypothetical patients who should have sought immediate medical care, ChatGPT Health told them they should stay home or wait to schedule a regular appointment with a doctor. 

The article, in the Guardian, said: "In one of the simulations, eight times out of 10 (84%), the platform sent a suffocating woman to a future appointment she would not live to see.... Meanwhile, 64.8% of completely safe individuals were told to seek immediate medical care."

For the study, published in the journal Nature Medicine, researchers created 60 realistic patient scenarios covering health conditions from mild illnesses to emergencies, then presented those scenarios to ChatGPT Health in various ways: changing the gender of the patient, sometimes providing test results, sometimes adding comments about what "friends" advised, etc. Three independent doctors reviewed each scenario and agreed on the level of care needed, based on clinical guidelines.

The study found that ChatGPT Health did well on textbook emergencies such as stroke and severe allergic reactions. But "'what worries me most,'" a doctor is quoted as saying in the article, "'is the false sense of security these systems create. If someone is told to wait 48 hours during an asthma attack or diabetic crisis, that reassurance could cost them their life.'”

Any number of health experts have extolled the potential for AI-based health advice, coupled with wearables and telemedicine, to revolutionize healthcare — providing care to the elderly and to people in rural areas, who would otherwise have difficulty getting access, while slowing the inexorable rise in healthcare costs. And I've bought in: Chunka Mui, Tim Andrews and I included a lengthy scenario about the potential for AI-based healthcare in our 2021 book, "A Brief History of a Perfect Future."

I still think the potential is there, too. As OpenAI, the developer of ChatGPT, told the Guardian, the app is updated and improved all the time, and I hope they keep charging ahead. (OpenAI also said it doesn't believe the study reflects how people actually use ChatGPT Health.)

But I also hope they are constantly checking for problems such as those identified in the study, and anyone else using AI in situations with major consequences should exercise similar care. That includes insurers, and not just in healthcare. As we feel our way toward using AI agents, we need to be very careful to not only vet them before putting them into production but to then supervise them — because they absolutely will make mistakes — and to keep improving them.

Cheers,

Paul

In the timeless words of Sun Tzu in The Art of War: "If you know the enemy and know yourself, you need not fear the result of a hundred battles." Today, in the battle against fraud in business and government programs, entity resolution—combined with knowledge graphs and geospatial analytics—serves as that ultimate weapon, akin to Excalibur, the legendary magical sword that could cut through anything.

When it comes to fighting fraud, it cuts through layers of deception, revealing hidden connections between people, businesses, transactions, and locations that fraudsters purposefully endeavor to keep obscured. By mapping out entities and resolving disparate records across dispersed systems to the real individuals and organizations behind them, investigators gain the clarity to validate transactions, expose invalid transactions, and dismantle fraudulent networks.

Fraud in government programs and business operations thrives in the shadows of fragmented data: mismatched names, shell companies, fake addresses, synthetic identities, and manipulated locations. Without a unified view, billions of dollars are lost annually to schemes like improper benefit claims, procurement kickbacks, subsidy abuse, "paper mills," and phantom vendor payments.

Entity resolution bridges these gaps, linking records across databases—names and addresses, tax filings, business registries, transaction logs, social media, and public records—to create a "360-degree" profile of every entity involved.

Entity Superpower — Unmasking the True Actors

At its heart, entity resolution determines when multiple records refer to the same real-world person, business, or location, despite variations in spelling, abbreviations, typos, or deliberate obfuscation. Advanced algorithms and machine learning handle the noise: "John A. Smith LLC" might resolve to the same entity as "JAS Enterprises" owned by "Jon Smith," especially when tied to shared addresses, phone numbers, or transaction patterns.

When integrated into knowledge graphs, these resolved entities form connected networks of relationships—ownership links, family ties, shared board members, or transaction flows. Adding the basics of address geocoding and geospatial analytics overlays physical reality: mapping addresses, proximity of claimed locations, or clustering of suspicious activities in specific regions. This data fusion transforms isolated data points into a battlefield looking glass that maps where fraud patterns emerge clearly.

Consider a classic red flag in government-funded programs: more licensed or funded daycares than the number of children in an area could possibly require. Entity resolution uncovers this by resolving provider records to actual owners and cross-referencing enrollment claims against demographic data. Knowledge graphs reveal networks of colluding owners registering multiple entities at the same address or funneling funds through connected shell companies. Geospatial views highlight unnatural concentrations—clusters of daycares in low-population rural zones or urban blocks with improbable child-to-provider ratios—signaling potential ghost operations or subsidy farming.

So, as with childcare, insurance companies may apply entity resolution to chiropractors, MRI facilities, and clinics, but in addition now the named insured, agent, claimant, and adjuster meld in with medical providers, equipment, legal staff, vendors, and others in the graph across any line of business. As lines are combined and companies join forces, this process can literally map trillions of dollars of historical premiums and claims that could influence real-time payments.

The King's Sword Trumps All Use Cases

Drawing from innovative applications across business and government using knowledge graphs for fraud detection, the combination of entity resolution, knowledge graphs, and geospatial tools exposes fraud across diverse domains:

• Government Benefit and Subsidy Fraud: In childcare subsidies, housing assistance, unemployment benefits, or agricultural grants, resolved entities expose operators claiming impossibly high volumes. Geospatial analysis flags unnatural provider distributions relative to demographics, while knowledge graphs uncover collusive networks funneling funds through connected shells or using stolen identities for enrollment claims.
• Procurement and Contract Fraud: Vendors often conceal conflicts via layered ownership or bid-rigging. Entity resolution connects bidders to officials' associates or hidden entities; geospatial overlays reveal fictitious delivery sites or illogical routing; graphs detect circular payments or anomalous bidding patterns indicative of corruption.
• Fake Business and Identity Schemes: Fraud rings create phantom companies for loans, grants, tax credits, or PPP-style programs. Resolution merges digital and physical footprints—such as mismatched websites/IPs with abandoned addresses—while geospatial clustering pinpoints registration hotspots tied to broader scams.
• Money Laundering and Illicit Flows: In trade-based or benefit-related schemes, resolved entities link actors across jurisdictions. Knowledge graphs map multi-hop transaction chains; geospatial tools visualize fund movements against claimed origins, exposing laundering through high-risk locations or mismatched geographies.
• Insurance Claims Fraud: In property insurance schemes, fraudsters stage incidents like water damage during homeowners' vacations, directing repairs to complicit restoration providers. Entity resolution links claimants, properties, and service providers across cases, revealing common identities or ownership ties; knowledge graphs highlight recurring patterns in damage types, timing, and vendor involvement; geospatial analytics maps claim locations against provider clusters, unmasking organized rings exploiting insureds and property owners.

In auto insurance, staged accidents generate multiple unrelated passengers all seeking medical treatment from the same provider and being represented by the same lawyer even though they themselves may live far apart and curiously are frequently unable to be located.

The schemes for various lines of casualty and property in auto, home, workers' compensation, and commercial insurance all are well mapped by the NICB (National Insurance Crime Bureau). And new schemes are emerging all the time — especially with the backing of transnational criminal organizations, but also with just everyday people getting creative with generative AI.

En Garde — the Industry Keeps Its Hand on the Hilt

As fraud schemes grow more sophisticated with digital mapping tools and global reach, entity resolution in knowledge graphs—enhanced by geospatial context—will only sharpen. Real-time monitoring, AI-driven anomaly detection, and dynamic mapping will make deception harder to sustain. The result? Interdiction of transactions. Faster and better recoveries. Frustrated, if not deterred, criminals. Lower premiums for insureds. Safeguarded public funds.

In the war on fraud, knowledge is power—but resolved, connected, and spatially aware knowledge is the key to victory. Like Excalibur drawn from the stone, we across these industries, companies, and public bodies draw data from our legacy and modern systems. This combination of data and technology empowers those who wield it to cut through illusion and restore justice.

Let's talk about something that's been hiding in plain sight in insurance and healthcare operations for the better part of three decades: You have no idea what your processes are actually doing.

I don't mean that as an insult. I mean it as a structural observation. You have dashboards—God, do you have dashboards. Gorgeous ones with KPI tiles and sparklines trending whichever way the builder needed them to trend. You have reporting teams producing decks for Monday standups—assemblies of data that's six weeks old, filtered through three layers of organizational telephone, and crafted—not maliciously, but inevitably—to support a story someone already believed.

What you mostly don't have is instrumentation. Real instrumentation. The kind that tells you, in something close to real time, what your core processes are producing, where they're breaking, and what that's costing you.

That gap is about to get much more expensive to ignore.

Process excellence folks will recognize DMAIC—Define, Measure, Analyze, Improve, Control. The problem is that in most operations, the M and the A have always been the expensive, politically fraught parts. So organizations Define—sometimes brilliantly—and then Jump. Straight to Improve. They hire consultants, run workshops, launch initiatives, celebrate launches. A year later, they do it again. That isn't improvement. It's expensive thrash—innovation theater in a process‑excellence costume.

Instrumentation was always theoretically worth it. It just never made it to the top of the list.

Enter AI, which changes this calculation in two ways—one a mandate, one a gift.

The mandate first, because it's the one that gets you fired.

You can't drop operational AI into a live process environment without knowing precisely what it's doing. AI systems in claims processing, prior authorization, utilization management—these make decisions at a speed and scale no human team can realistically audit afterward. If you don't have instrumentation showing, in near‑real‑time, what your models are producing, where they're drifting, and where edge cases are piling up into systematic errors, you'll have a very bad day. Possibly a regulatory very bad day. Possibly a front‑page very bad day.

Operational AI forces the instrumentation conversation in a way Six Sigma consultants never could.

Now the gift.

AI also makes instrumentation cheaper and easier than it's ever been. Process‑mining tools can map your actual workflows—not the idealized Visio diagram, but what's really happening—by reading keystrokes, logs and system events that already exist. Natural language processing (NLP) can monitor unstructured outputs: call transcripts, clinical notes, adjuster comments, member complaints. Modern data pipelines can connect legacy systems in a fraction of the former time and cost. All without creating risk or dependencies.

By instrumenting your operation for AI, you end up using AI to measure what you should have been measuring all along. The mandate and the gift are the same. You don't get the AI transformation without building the measurement infrastructure—and once you've built it, you finally have something most organizations have never possessed: a real‑time picture of their own operations.

The counterintuitive part nobody talks about: people assume a fully instrumented, heavily automated operation becomes robotic. Soulless.

The opposite is true.

When 80% of your operation runs smoothly—instrumented, measured, automated, in control—something remarkable happens to your meetings. The variance archaeology, the defensive explaining, the "why did this metric move?" inquisitions—all move into dashboards that don't need a room full of people to interpret. What's left in your daily standup are exceptions. Real exceptions. The claim that fell outside every parameter. The member experience that defied categorization.

Exceptions are where operations learn. They're where customer‑service stories live—the quietly devastating and the genuinely remarkable—and those stories, surfaced in a room of engaged humans, are where innovation happens. Not in workshops or hackathons, but in noticing an exception, connecting it to context, and realizing it points to something structural.

The daily meeting becomes tactical again—focused on real issues, resolved quickly, without drifting into philosophical fog. Strategy moves to the quarterly business review, where it belongs. Mixing the daily and the quarterly is how organizations end up doing neither well.

The even better news is that a genuinely well‑run operation—one that knows what it's doing, measures what matters, and improves based on evidence—can deliver on a real mission. Instrumentation isn't separate from culture; it's the infrastructure culture runs on.

The more automated your operation becomes, the more human it can afford to be.

The instrumentation imperative is real, and AI is making it urgent. The organizations that win will be the ones that treat it not as compliance, but as what they should have built 20 years ago—finally achievable, finally affordable, and harder to ignore every quarter they wait.

Paul Carroll

AI is everywhere in insurance right now. Where do you see it being used especially well?

Dr. Michael Bewley

One mature application is the use of something called supervised machine learning, for aerial imagery. The application provides a way of getting reliable recognition of objects and images, which can be really informative about a property. Then you can use what you see in trusted frameworks. You know, given the roof had large patches of rusting or missing shingles or a hole in it before the event, what's the likelihood of damage in the event? That can be modeled in a pretty clean way.

But there's a whole spectrum of AI from really quantifiable, reliable, and well-understood systems all the way through to things where it's all about creativity. You throw in an idea, and it comes up with some more ideas. 

Even traditional risk modeling can be seen as AI. You're trying to predict the likelihood of claims.

Paul Carroll

What are the risks associated with using AI?

Dr. Michael Bewley

You've got to get the newly available data and realize it's amazing but then apply it carefully, because all data comes with uncertainty. Even if we're really confident it's a solar panel on that roof, we'll tell customers we're 98% confident. There's a 2% chance we're wrong, and saying so allows insurers to treat the data in a more nuanced way.

Paul Carroll

There's growing pushback against AI-based property assessments. People are told they have a roof problem through AI and aerial surveillance, and while they may acknowledge the issue is real, they resist being charged based on that information. How do you address this customer trust challenge?

Dr. Michael Bewley

That we can determine a roof's condition remotely is really valuable—not just to the insurer, but to the insured. Not many people climb on their roof on a regular basis. The fact that we can not only say there's an issue with that roof, but we can show the image it comes from is really important.

If we can tell someone their roof is damaged, they can fix it. They can reduce their risk, and that's in everyone's interest.

Paul Carroll

Organizations are dealing with information in countless different forms—one insurance system had 37 ways that San Francisco was described, from San Francisco to San Fran, SF, Frisco and so on. Is this data uncertainty part of the reason property-related decisions are still so difficult to make?

Dr. Michael Bewley

Just having so much more data today doesn't necessarily make for good decisions in and of itself. There are so many questionable sources of information out there, and there are so many sources where it's unclear how accurate they are because you can't actually see the provenance. It's very difficult to ascribe a level of trust.

This is why we've hinged our whole strategy on aerial imagery. We bring in third-party data and other information, but the core is what your eyes can see.

Insurers are being bombarded by a huge range of information from different vendors and open information out there on the Web. So we're very particular about how we form our information, and we make that transparent to the user. Every bit of data that we serve up in our APIs comes with a link so you can go and look at the photo. 

It's well-articulated information that matters. Volume can actually be a detractor because you get lost in the noise.

Paul Carroll

The insurance industry has historically moved slowly, but in catastrophe response, speed is critical. Where do we speed things up?

Dr. Michael Bewley

The challenge is that a catastrophe is a continually unraveling scenario. It's not just that the cat event occurs, then we're done, and we all move on. The hurricane makes landfall, properties get damaged, the storm keeps moving, further events occur, there are recovery efforts, and so on. So while speed is good, clarity is important, as well. 

If there's an event that we're going to capture with our cameras, we'll get a plane up in the air as soon as it's safe. As soon as we capture some valid imagery, we turn it around as fast as we can, using AI. In Hurricane Milton, I think we flew over 100 flights because there were so many things going on—the weather changes, what's going on on the ground changes.

Paul Carroll

Would you talk a bit more about how insurance can move from the traditional repair-and-replace model to a Predict & Prevent approach?

Dr. Michael Bewley

That's a great question. If we step back from the catastrophe-specific discussion, our regular capture program covers most well-populated areas multiple times a year. We’ve done this for a decade now in the U.S. and 18 years in Australia.

The regular uptake of imagery, year in, year out, shows you where things are today and where they've been historically, and then captures an event in that context. A really good example is our new roof edge product. We've run AI on stupendous quantities of imagery. We've looked at our full imagery archive in the U.S. and run every single house on every single historical date to work out when a new roof got put in. If an event is coming up, you can start to feed that into an understanding of whether the roof is getting to end of life anyway, so maybe it's time to replace it. Maybe that reduces the risk. You can have a mature discussion between the insured and the insurer about that. 

The exact same imagery is being used by insurers, by local governments, by construction, by town planning, by environmental groups, by so many different sorts of people. So they can have discussions about how to remediate the risks on a property before an event happens. We can talk about how we plan towns better. It's wonderful if we can all look at that same source of truth.

Paul Carroll 

What is one challenge you'd like to offer to insurers about their assumptions on property risk? What are they missing that they should understand?

Dr. Michael Bewley

I think the challenge is really for them to understand that there are new. high-quality sources of information available. They may be used to doing things a certain way with limited information, so they have to understand the incoming information and make good use of it.

In the AI space, the challenge is sifting the signal from the noise. There is genuinely a bunch of AI stuff, particularly the stuff that's in the media a lot, that one needs to treat very carefully. All the large language models and Gen AI imagery stuff—there is a place for that in insurance, but it's different from the more tried-and-tested machine learning approaches, and we have to weave that in carefully. It's very important to understand the full tapestry of AI solutions that there are and not to get them muddled up. 

Gen AI opened up a new world. It is absolutely revolutionary. I think it's on the level of the internet being invented or the personal computer. So you definitely don't want to sit by and say, "Well, I'll wait and see what happens," or "This one's not for me." You've got to get involved.

But as with the personal computer and the internet coming online, there's uncertainty about how to use it. There's uncertainty about what the impact will be. You just have to get in there and get involved. But you have to do it with wisdom and care.

Paul Carroll

Yeah, I think we've just scratched the surface. This is quite a ride we're on.

About Dr. Michael Bewley

Dr. Michael Bewley’s passion for AI began in 2007. Graduating with degrees in electrical engineering and physics (University of Sydney), he received the University Medal for using machine learning (ML) on brain scans to detect Alzheimer’s disease. He joined Cochlear to work on implantable hearing solutions, also implementing its first customer-use product analytics.  A sea-change led to a PhD program at the Australian Centre for Field Robotics, using ML to interpret sea-floor imagery from autonomous submersible surveys. He also established a data science team as Lead Data Scientist at the Commonwealth Bank.  Mike joined Nearmap in 2017 and is now VP of AI & Computer Vision, leading the development of AI technology, applying petabyte-scale deep learning on geospatial imagery and AI data sets.

Subpoenas are a routine part of claims investigations, coverage disputes and regulatory inquiries. What isn't routine is the pace at which they're arriving.

A new analysis from Wolters Kluwer CT found that U.S. subpoena volumes reached 498,000 in 2025, with growth accelerating year-over-year. After a brief 3% dip in 2020, volumes have climbed every year since, growing 13% in 2023, 10% in 2024 and 8% in 2025. Insurance is absorbing more of that increase than any other sector.

Insurance-related subpoenas grew 65% between 2019 and 2025, making them the fastest-growing category in the data. Roughly 80% of that activity is concentrated in California, Florida, Georgia and Texas. Each of these jurisdictions has its own combination of regulatory activity, litigation trends and natural disaster exposure fueling the increase.

Florida saw the sharpest increase of any state, with volumes up 86% since 2019. Hurricane claim investigations, growing demand for insurance-related records and a wave of litigation filed ahead of major tort reform measures all contributed. Much of this reflects heightened scrutiny from state regulators examining how insurers handle claims in disaster-affected areas, which generates a downstream surge in records requests and legal process activity. California volumes rose 54%, driven by insurance coverage disputes, surplus line insurer activity and new privacy compliance requirements. California's evolving regulatory landscape, particularly around data access and consumer protection, has expanded the scope of what gets subpoenaed and how quickly insurers are expected to respond.

The implications extend beyond claims departments into insurance legal operations and compliance.

How intake processes need to change

For most insurers, subpoena intake and response processes were built for a different volume environment. Many still rely on manual workflows to receive, triage and route incoming legal documents across departments and jurisdictions. Gradual increases are manageable. A 65% jump in six years exposes the limits of processes that were never designed for this pace.

Missed response deadlines create legal exposure. Misdirected documents delay claims resolution. Inconsistent handling across state lines introduces compliance risk, particularly for multi-state insurers navigating different procedural requirements in each jurisdiction. The operational cost of getting it wrong is compounding as volumes climb.

Jurisdictional complexity adds to the burden

The geographic concentration of subpoena growth creates a particular challenge for insurers that operate across states. Multi-state insurers are managing higher volumes under different rules, different timelines and different regulatory expectations in each jurisdiction.

With roughly 80% of insurance-related subpoena activity concentrated in four states, organizations with significant exposure in Florida, California, Georgia and Texas face a disproportionate operational burden. The resource allocation models and response frameworks that worked five years ago may no longer be adequate for today's volume and complexity.

What insurers should do now

The subpoena data points to a broader reality about litigation complexity that extends beyond any single sector. Regulatory scrutiny is increasing, data access expectations are broadening and legal activity in key sectors is accelerating. These are structural trends, not temporary spikes.

Insurers managing legal process intake through fragmented, manual systems are absorbing unnecessary risk. The organizations best positioned to handle this environment are the ones treating legal process management as an operational discipline rather than an administrative afterthought.

That means evaluating how subpoenas and other legal documents are received, tracked and routed across the organization. It means understanding jurisdictional requirements at a granular level and building response protocols that account for the specific procedural obligations in high-volume states. Subpoena volume trends also signal where litigation and regulatory activity are heading, which should inform how insurers staff and structure their legal process operations.

If these trends hold, the gap between current legal process volumes and most insurers' capacity to manage them will only widen. The question for insurers is whether their legal operations are built for the volume they're handling today or the volume they were handling five years ago.

For years, even decades, senior leaders in the insurance industry have pursued the goal of fully digitized claims operations. The business case was especially strong for straightforward property and casualty claims, where high volumes and repeatable patterns made automation attractive. Still, carriers across all lines of business saw the potential benefits of streamlining workflows. The logic was simple. If insurers could automatically capture the right data, use claims processing automation to handle routine steps, and speed payouts, operating costs would decline, and customer satisfaction would improve.

Today, for many insurers, that vision is no longer theoretical. With the help of claim management automation solutions, routine claims can now move through the system with limited manual intervention. Costs have come down, timelines have shortened, and straightforward claims are often resolved faster than ever before.

But this progress raises a new question. Now that efficiency has improved, what comes next?

Why This Conversation Matters Now

For many years, claims transformation was defined by speed. Insurers focused on faster first notice of loss, assessment, adjudication, and payout. Speed became the main indicator of progress.

Speed still matters. Delays create financial strain for customers and reputational strain for insurers. But in 2026, speed alone is no longer sufficient.

1. The Need for Fairness and Defensibility

Insurance companies promise more than financial payment. They promise fair treatment. When a customer files a claim, they are often stressed or confused. In that moment, how the claim is handled matters as much as the final settlement. A delayed response, unclear explanation, inconsistent decision, or weak documentation can quickly escalate into a bad-faith allegation. Once that happens, legal costs rise, and reputational damage follows.

This is where automated claims processing insurance platforms are gaining attention. Beyond efficiency, they establish clearer documentation and consistent workflows.

They also create traceable decision pathways with well-articulated audit trails. Such a lucid and transparent structure enables insurers to demonstrate that claims were handled judiciously and in good faith.

2. Rising Complexity and Fraud

Claims complexity is also increasing. CAT events are more frequent and destructive. Fraud schemes are more coordinated. Regulatory oversight is more exacting. Each decision may be reviewed months or even years later.

Fraud alone presents enormous pressure. Deloitte estimates suggest that roughly 10% of property and casualty claims are fraudulent, contributing to approximately $122 billion in annual losses. Deloitte also projects that by implementing AI-driven technologies across the claims life cycle and integrating real-time analysis from multiple data sources, P&C insurers could reduce fraudulent claims and save between $80 billion and $160 billion by 2032.

Modern insurance claims automation solutions help detect suspicious patterns at an early stage. High-risk claims are then routed for deeper scrutiny. This enables insurers to mitigate fraudulent activity. It also shields legitimate policyholders from the downstream repercussions of deceit.

3. Changing Risk Profiles Due to Workforce Strain

While claims complexity rises, the workforce is under strain. Many experienced claims professionals have retired. Institutional memory has thinned. Newer adjusters manage heavy caseloads with less experience. This creates uneven judgment and operational fragility.

With more advanced, affordable AI-based claims management automation solutions available, insurers have an opportunity to rethink the role of claims altogether. Instead of viewing claims purely as a cost center, forward-looking carriers are exploring how smarter, data-driven claims operations can create value. This includes improving loss ratios through better fraud detection and prevention, offering more personalized claims experiences, and even using insights from claims data to reduce future losses.

Automation and algorithmic decision-making are now common. Systems evaluate, approve, flag, and sometimes deny claims with limited human involvement. These tools increase efficiency. They also raise questions about accountability, bias, and explainability.

The central question has shifted. The industry must now ask not how fast claims can move, but how intelligently they can be handled.

The New Era of Claims Management

The future of claims processing is not about moving faster through workflows. It is about making better decisions at every step. Here are the core characteristics of the future of claims:

I. Balance

Smarter claims processing balances speed with accuracy. It balances automation with human judgment. It also balances efficiency with trust.

II. Fairness

A claim processed quickly but incorrectly creates rework, complaints, and litigation. An automated claim without context can harm a vulnerable customer. A decision issued without a clear rationale can invite regulatory scrutiny. The resulting fairness and transparency instills greater trust in the insurer-insured relationship.

III. Quality

Claims performance must be evaluated through decision quality. Cycle time and cost remain important, but they are incomplete measures. A high-quality decision is consistent, fair, traceable, and defensible.

Modern claims processing solutions should therefore be judged not only by how quickly files move, but by how reliably they withstand complaints, audits, and disputes.

Why Speed-First Models Are Breaking Down

Speed-first models were built for a different era. They assumed predictable claims, stable risk patterns, and clean data at intake. That environment no longer exists.

• Built for a Simpler Environment

Speed-first claims models were built for predictability. They assumed standard patterns and limited variation. Claims were treated like transactions moving down a straight pipeline.

That assumption no longer holds.

Today's claims are more varied. Policies are more complex. Weather-related losses are larger and less predictable. Fraud tactics are more organized. What once worked for routine cases now struggles under real-world pressure.

• Weak Intake Leads to Faster Mistakes

When intake data is incomplete and claims processing automation pushes the file forward anyway, errors spread quickly. Missing documents, incorrect coding, or misread policy terms can move through the system without being caught.

Automated claims processing insurance systems do not fix weak inputs on their own. They can magnify them. An improper denial can move just as quickly as a correct approval. When that happens, complaints rise. Rework increases. Legal risk grows.

Strong claims processing solutions must therefore focus on data accuracy at the start, not just speed at the end.

• Over-Automation Reduces Judgment

Over-automation also creates rigidity. Rule-driven systems work well for simple claims. A broken windshield or minor water leak may follow a clear path.

But many claims are not simple. A severe storm loss, a multi-party liability dispute, or a policyholder in financial distress requires context. It requires judgment. Claim management automation solutions should guide these cases, not force them into narrow rules.

Insurance claims automation solutions must be able to flag unusual patterns and route them for review. If everything is treated the same, fairness suffers.

• Explainability and Trust Are at Risk

Explainability is another weakness of speed-first models. A rapid decision without a clear explanation erodes trust. Customers may feel ignored. Regulators may question whether similar cases are handled the same way. Leaders may struggle to defend outcomes during audits.

Clear documentation matters. Claims processing automation should record what was reviewed, what rules were applied, and why a decision was made. Without that record, even a correct decision looks careless.

•  Automation Without Intelligence

When it comes to claims processing, the problem is not automation itself. Claims processing automation can reduce manual errors and improve consistency. Automated claims processing insurance systems can shorten timelines and improve service.

The problem is automation without thought. It offers speed without review, and establishes rules without room for context.

The next stage of claims modernization must combine structure with judgment. Automation should support sound decisions, not replace them.

What Smarter Claims Really Means

Smarter claims processing has a practical definition. It means using technology to support sound judgment rather than replace it.

AI-driven automated claims processing systems can:

• Extract and verify data from documents
• Compare claim details against policy terms
• Detect fraud patterns across large datasets
• Prioritize claims by complexity and risk
• Route sensitive cases for human review
• Provide clear documentation of every step taken

This does not eliminate the insurer's legal duty to act reasonably. It helps fulfill that duty more consistently.

Smarter claims automation systems integrate  integrate policy data, claimant history, prior outcomes, and external signals before guiding decisions. Straightforward claims move quickly. Complex or high-risk claims receive deeper review.

Learning is embedded in the system. Complaints, reversals, litigation outcomes, and regulatory findings feed back into decision support models. Over time, the system becomes more refined and less erratic.

Even modest improvements matter. Best-in-class insurers applying AI in specific domains have already achieved measurable results, including a 3% to 5% improvement in claims accuracy. That may seem small, but at scale it can mean thousands fewer disputes.

The Shift from Workflow Engines to Decision Engines

Traditional claims platforms functioned as workflow engines. They moved files from one predefined step to the next. The focus was on process efficiency.

Modern claims capabilities are evolving into decision engines.

Instead of simply pushing tasks forward, decision engines evaluate context and risk in real time. They determine whether a claim should be automated, referred, or escalated. They assess gradients of complexity rather than forcing uniform treatment.

In a workflow model, success is defined by movement. However, in a decision model, success is defined by the integrity of the outcome.

This structural shift strengthens defensibility when decisions are later challenged.

How Trust Has Become the New KPI

As automation deepens, trust becomes central.

For starters, customers want to understand why their claim was approved, adjusted, or denied. As such, transparency is no longer optional.

On the other hand, regulators expect traceability. They want audit trails that show how data flowed through systems and how conclusions were reached.

Finally, executives expect risk control. They want assurance that automation does not introduce hidden bias or unpredictable exposure.

Trust can be measured through:

• Lower complaint volumes
• Fewer bad-faith allegations
• Reduced litigation frequency
• Consistent audit outcomes

Smarter claims systems embed traceability and governance into the decision path itself. They generate documentation in real time rather than reconstructing it after disputes arise.

It is vital to note that trust is not built on speed. It is built on clarity and consistency.

What CIOs Need to Focus on Now

For CIOs, smarter claims processing is not just a technology upgrade. It is a capability shift. Here's what they should focus on:

• Claims should be treated as a decision system. Investments must support contextual insight, structured judgment, and adaptive routing.
• Data quality must be strengthened at intake. Weak upstream data produces fragile downstream outcomes.
• Human oversight needs to be intentional. It cannot be perfunctory or symbolic. Thresholds for escalation must be clearly defined. Mechanisms for override and structured pathways for review should remain controlled and unambiguous.
• Governance is not optional. It is foundational. Explainability, audit trails, and bias monitoring cannot be treated as incidental add-ons or postscript considerations. They must be embedded from the outset.
• Metrics need constant recalibration. Static scorecards will not suffice. Beyond cycle time and cost efficiency, insurers should track decision consistency and complaint frequency. They must also monitor litigation exposure and fraud-detection efficacy with greater granularity.

All in all, claims modernization is not about acceleration alone. It is about discernment and prudent judgment. Speed matters, of course. But sagacity matters more.

The Bottom Line

Insurance companies promise fair treatment, not just fast payment. In a volatile and heavily scrutinized environment, that promise must be defensible and demonstrable.

The future of claims will continue to value efficiency. Its defining attribute, however, will be intelligence and calibrated reasoning.

Insurers that prioritize decision quality alongside speed will be better positioned for long-term resilience. They will reduce bad-faith exposure and manage fraud risk with greater dexterity. They will also sustain regulatory confidence and preserve customer trust.

The next phase of transformation will hinge on responsible claims stewardship. Ethical automation, explicit oversight, and equitable decision-making will be indispensable. Insurers that combine claims processing automation with transparency and robust governance will not merely control costs. They will fortify customer trust and cultivate enduring loyalty.