Download

How to Manage Teen Drivers' Risks

Adding a teen driver can nearly triple premiums, but proper coverage coordination and early planning help families manage risk.

Insurance

A teenager's first driver's license represents freedom, responsibility and, for parents, more than a little anxiety. For insurers and independent agents, it also introduces one of personal auto's most challenging exposures: a driver with little experience, no established record and a statistically higher likelihood of a crash.

The financial effect can be dramatic. Bankrate data cited by CNBC estimate that adding a 16-year-old to a family "full-coverage" policy can increase the average annual premium from $2,671 to $5,910. The human cost matters more. AAA reports that 2,636 people were killed in crashes involving teen drivers in 2024, underscoring that the consequences extend beyond young drivers to their passengers and others sharing the road.

For most families, insurers address this exposure within the personal auto policy rather than through a separate teen-driver product. They adapt through inexperienced-operator classifications, vehicle-assignment rules, underwriting requirements, and pricing that reflects years licensed, vehicle type, garaging, mileage, and use.

What Agents Should Do

For agents, the process should begin when the teenager receives a learner's permit, not after the first solo trip. The first obligation is accurate disclosure of everyone the application or policy requires to be identified, including household residents and customary operators. How a disclosed driver is treated for rating purposes is a separate question governed by the insurer's filed rules. This early conversation also provides an opportunity to discuss formal training. In some states, waiting until age 18 may eliminate a mandatory driver-education requirement, but it does not make the newly licensed driver experienced and may mean missing an available training discount.

That distinction becomes especially important when parents live separately. The ISO Personal Auto Policy uses "family member" for a person related to the named insured by blood, marriage or adoption who resides in the named insured's household. The Massachusetts Auto Policy uses the term "household member," with a substantially similar definition. Neither definition designates a single primary residence or makes the address on a license or registration controlling. Depending on the child's actual living arrangement, the child may qualify under policies issued to both parents. Agents should therefore review both households, both parents' auto policies and umbrellas, and every vehicle the teen may operate.

Ownership creates another potential trap. Some families place the policy in the child's name while the parent retains legal title to the vehicle. A carrier may accept that arrangement, but acceptance does not change the policy definitions. Being listed as an operator does not necessarily give the parent the rights or insured status of the named insured. The mismatch can affect liability arising from ownership, permission for other drivers, automatic coverage for a replacement vehicle, physical-damage proceeds and umbrella protection.

Coverage Vs. Price

Coverage should receive at least as much attention as price. Liability and uninsured or underinsured motorist limits, physical-damage coverage, deductibles and the family's personal umbrella should all be considered. A resident teen relative generally qualifies as an insured under the ISO personal umbrella, but forms vary and the underlying auto insurance must be coordinated.

Affordability still matters. Depending on the carrier and state, savings may come from good-student or low-mileage discounts, completion of approved driver education or advanced crash-prevention training, and telematics programs that evaluate speeding, hard braking, mileage, time of day and cellphone distraction and provide financial incentives for safer driving.

Insurance cannot eliminate the risk that comes with a new driver. But thoughtful underwriting, properly coordinated coverage, safety-focused incentives and an informed independent agent can help families manage it—giving teenagers room to gain experience and parents a little more peace of mind.

3 Keys for an Accurate View of Catastrophe RIsks

Insurers face a $171 billion global catastrophe benchmark as modeling advances reveal risks far beyond what recent loss experience suggests.

loss

$117 billion: the insured average annual loss (AAL) estimated for natural catastrophe events in the United States, according to Verisk's 2026 Global Modeled Catastrophe Losses Report. This figure represents roughly two-thirds of the $171 billion global AAL benchmark. The AAL is not a forecast for next year; it is the level of loss the insurance industry should be prepared to absorb over the long term. North America remains the dominant contributor, accounting for $124 billion of the global total.

A benchmark built for today's risk environment

Global insured losses have topped $100 billion for each of the past six years. Since the report's first edition in 2012, estimated global modeled insured average annual losses have increased from $59 billion (in 2011 USD) to $171 billion (in 2025 USD). The increase reflects several factors, including exposure growth, expanded model coverage, advances in catastrophe science, and continuing refinement of the industry's view of risk. The 2026 AAL benchmark also reflects the latest generation of model updates, so it captures current science rather than relying on the assumptions of the last decade.

Why the average is only part of the story

The insured AAL is a long-run expectation, not a prediction. Especially in the U.S., frequent perils such as severe thunderstorms, wildfires, inland floods, and winter storms drive the average, while severe perils, including hurricanes and earthquakes, drive the extremes. Losses in any single year can land far from the mean, which is why the full range of modeled outcomes matters more than any one number.

The difference between average and extreme outcomes illustrates why insurers focus on more than the mean. Against the $171 billion global AAL, the same models estimate a 100-year industry loss approaching $477 billion and a 250-year loss exceeding $600 billion. Catastrophe models are designed to look beyond historical experience and quantify plausible events the industry has not yet experienced.

A single year of lower loss activity doesn't mean lower risk

Catastrophe losses were below average in 2025, but a below-average year reflects a favorable draw, not a change in the risk landscape. Despite the absence of a landfalling hurricane in the continental United States, global insured catastrophe losses still exceeded $100 billion. Wildfires and severe thunderstorms demonstrated how the accumulation of frequency-peril losses can generate significant industry losses even without a headline catastrophe event. The first half of 2026 further illustrated the point. The U.S. p&c industry's combined ratio improved to 92.7 from 96.5 a year earlier, and its estimated net underwriting gain nearly tripled to $31.7 billion. Those results show where the industry landed on the curve this year, not a shift in the underlying risk. The $117 billion expectation for the U.S. doesn't shrink because the first half of the year was kind.

A favorable year can improve financial results, but it does not alter the underlying distribution of catastrophe risk that insurers must manage.

Better to be prepared than surprised

For risk leaders, preparation starts with the AAL, the anchor for catastrophe risk appetite and capital planning and the number with the deepest data behind it. Catastrophe models rely on increasingly detailed exposure information, including building characteristics, reconstruction costs, and claims experience. The quality and currency of that data influence how insurers evaluate risk appetite, reinsurance needs, reserving assumptions, and capital allocation decisions. As exposures evolve and catastrophe science advances, maintaining an accurate view of risk becomes increasingly important.

Three considerations stand out in an environment where catastrophe losses regularly exceed historical expectations. First, evaluate risk appetite against the full range of modeled outcomes rather than recent loss experience alone; the tail of the distribution is often what places the greatest strain on balance sheets. Second, consider how reinsurance programs respond to the perils that drive extreme loss scenarios, not just the events dominating recent headlines. Third, reassess capital assumptions when exposure profiles or catastrophe models change materially, rather than relying solely on periodic review cycles. The purpose of catastrophe modeling is to look beyond what happened this year and quantify what could happen across thousands of plausible future scenarios. The industry's challenge is not preparing for the year it just experienced, but for the years it has not yet seen.

Sources

2026 Global Modeled Catastrophe Losses Report: https://www.verisk.com/resources/campaigns/modeling-insured-catastrophe-losses-a-global-perspective/

1H 2026 underwriting results: https://www.verisk.com/company/newsroom/


Jay Guin

Profile picture for user JayGuin

Jay Guin

Dr. Jay Guin is the executive vice president and chief research officer for Verisk Catastrophe and Risk Solutions. 

With more than 25 years of experience in catastrophe modeling and risk analytics, he oversees research initiatives that support risk assessment across a wide range of natural and emerging perils.

What Job Descriptions Show About AI Adoption

Insurers are hiring AI builders, but descriptions for operational roles don't yet reflect how claims and underwriting work will change.

Insurance

Investments in AI are beginning to change the requirements placed on the insurance workforce. As AI becomes embedded in daily operations, employees will need to interpret system-generated insights, manage exceptions, document AI-assisted decisions, and apply judgment in situations where automated tools lack sufficient context.

Shift recently examined 100 job postings from leading U.S. insurance organizations to understand how this transition is appearing in the hiring market. The findings show substantial demand for the people building AI capabilities. Seventy-two percent of AI, analytics, product, and technology roles explicitly referenced AI. Among operating roles in claims, underwriting, SIU, and subrogation, the figure was 6%.

The results suggest that insurers have entered the next stage of AI adoption. Technical teams are building the infrastructure and capabilities. The work now extends to preparing the broader organization to use those capabilities effectively.

Job descriptions reveal how work is evolving

Job descriptions serve several purposes. They attract candidates, define current responsibilities, and signal which capabilities an organization expects to matter in the future. Viewed collectively, they can also show how an industry is preparing for changes in technology and operating models.

The hiring patterns in Shift's research suggest that insurers have concentrated much of their visible AI hiring within teams responsible for building, implementing and supporting the technology. These roles require specialized expertise in data science, product development, analytics and engineering.

The impact of those investments will eventually extend across the business. Claims professionals, underwriters, investigators and other operational teams will increasingly use AI-enabled tools as part of their daily work. Their responsibilities will evolve as those systems become more deeply integrated into decisions and workflows.

That transition is beginning to appear in the hiring landscape. Operational job descriptions continue to emphasize core insurance capabilities such as judgment, communication, and customer service. Far fewer describe responsibilities involving AI-generated insights, exception management, documentation of AI-assisted decisions, or the responsible use of automated recommendations.

Preparing for AI-Enabled Insurance Work

Insurance expertise remains essential. Human judgment is still one of the industry's most valued skills: nearly eight in 10 job postings analyzed referenced judgment, discretion, complex decision-making, investigation, negotiation or problem-solving. More than three-quarters of operating-role postings emphasized these capabilities.

AI will change how that expertise is applied. As automated systems take on more routine tasks, employees can devote more time to complex claims, disputed liability, ambiguous policy language, and sensitive customer situations. These scenarios require context, experience, and decision-making that cannot be reduced to a system output.

The findings show that current job descriptions value judgment, although they rarely describe it in the context of AI-enabled work. Few operating-role postings mention AI-assisted decision support, and none reference the responsible use of AI.

Insurers can begin to define these capabilities more precisely. Judgment includes assessing whether an AI-supported recommendation makes sense in context. Investigation increasingly includes determining what additional information is needed when a system flags a claim, policy, provider or network. Communication includes explaining a decision clearly and empathetically when digital tools have informed the process.

In this environment, AI can increase the value of insurance expertise by helping professionals apply it more consistently and at greater scale.

The entry-level postings in the sample offer an early indication of how this development may occur. AI references were uncommon across claims roles, though some trainee positions specifically mentioned a willingness to learn or receive training in AI.

This language suggests that some insurers may plan to develop AI fluency through onboarding, internal learning and day-to-day use of AI-enabled tools. Such an approach could be especially valuable in an industry facing an experience and talent gap. AI can help newer employees access institutional knowledge, identify relevant information and develop sound decision-making practices more quickly.

The job-posting data cannot show how widespread these training programs are. It does, however, raise an important workforce question: how should insurers combine hiring, upskilling and workflow design to prepare operating teams for AI-enabled work?

Bringing Responsible AI into Operational Roles

Responsible AI appears rarely in insurance job descriptions. References to explainability, governance, transparency, and responsible use occur occasionally in analytics and technical roles. None of the operating-role postings analyzed included this language.

As AI becomes part of everyday insurance decisions, governance also becomes an operational capability. Employees using AI-supported systems may need to recognize when a recommendation warrants further review, determine when additional evidence is required, document how AI informed a decision, and escalate cases when a system lacks sufficient confidence or context.

Human-in-the-loop design, auditability and operational controls begin at the technology level, then continue through the workflow. Their effectiveness depends on the people applying them in real situations.

Many insurers may currently address responsible AI through technical, legal, risk or compliance teams. Greater adoption will create a need to translate those principles into practical expectations for claims professionals, underwriters, investigators and managers.

The Next Phase

The hiring data shows that insurers are investing in the technical talent needed to build and deploy AI. The next phase involves integrating those capabilities into the work of claims professionals, underwriters, investigators, and other operating teams.

Traditional insurance expertise will remain the foundation of these roles. Employees will apply that expertise in an environment shaped by AI-assisted insights, automated workflows and increasingly sophisticated decision-support systems. Job descriptions, onboarding, training and leadership expectations can help define how that work should be performed.

The value of AI will ultimately depend on the quality of the decisions it helps people make. Insurers that combine strong technology with experienced, well-prepared employees will be better positioned to improve consistency, manage complex cases and serve customers effectively.

AI transformation is therefore also a workforce transformation. Its success will depend on helping insurance professionals use new capabilities with judgment, confidence and accountability.


Éric Sibony

Profile picture for user EricSibony

Éric Sibony

Éric Sibony is the co-founder of Shift Technology.

He holds a degree in applied mathematics from École Polytechnique, as well as a master’s degree in probability and finance jointly awarded by École Polytechnique and Université Paris VI. He also holds a PhD in mathematical modeling and machine learning from Télécom ParisTech.

How AI Can Orchestrate Claims Workflow

Workflow orchestration in workers' comp can use language models to automate routine claim responses while keeping critical decisions with adjusters.

Workers comp

When a representation letter lands on a workers' comp claim, an experienced adjuster does roughly the same things every time. They acknowledge the letter, reach out to the claimant's attorney, adjust reserves now that an attorney is involved, and update the claim summary. Ask 10 adjusters for the sequence and you'll get the same list in slightly different orders, with a step or two of local flavor.

The same holds when a return-to-work note comes in, full duty or with restrictions, or a demand package. It even holds for silence: a claimant who hasn't heard from anyone in three weeks is its own trigger, with its own response. These sequences repeat from file to file, and mostly from shop to shop. An adjuster with 15 years of experience has run some of them a thousand times, and could run them half asleep.

Workflow orchestration gets defined in a lot of ways. If we keep it really simple, it comes down to this: when something needs to happen on a claim, the right response follows.

Done well, the routine response happens on its own, and the exceptions land with a person quickly. The way to get there is a system that learns the actions taken on past claims, when they were taken and what they led to, and then executes them automatically. In other words, it's like a new hire on the claims floor, one who happens to have read every claim the company ever closed and can be trusted with the motions everyone already agrees on. The calls that matter stay with whoever owns the file. And like any new hire, the system gets brought along in phases. None of this was practical before, because most of a claim file is prose. But language models can finally read it.

Where to Start

The place to start is simple, getting the work itself into view. What's happening on each file and what the adjuster is doing about it, gathered in one place instead of spread across the diary, the inboxes, and the core system. Today the diary is the closest thing to that picture. A file comes up because 60 days passed, not because anything happened on it, and the first half hour goes to finding out whether anything did.

An overview like that is worth building before a single routine runs on its own. It shows what changed overnight and what's waiting on a decision, and it starts the adjuster's morning at the right files. It also means the state of a file no longer lives only in the adjuster's head, so nothing stalls when someone is out for two weeks, and a handoff doesn't start from zero. And it gives a claims operation something it has never really had: a clear picture of its own work. Everything else, the suggestions and eventually the actions, builds on that clarity, because work can only be handed to a system once it can be seen. The phases are how that picture gets built.

Phase One: Analyze

Every one of these sequences starts from a trigger. Sometimes the trigger is an event, like a letter arriving or a claimant going quiet. Sometimes it's a scheduled action: a state deadline on the first payment, or a follow-up that was promised and hasn't happened yet. The first phase is about proving those triggers get caught. The system doesn't act on anything yet. It reads everything that arrives and reports what it sees: what changed on the file since yesterday, and which of those changes should set a sequence in motion.

The same reading, pointed backward, is where the insights start. Once the files are readable, questions that used to require a file review project can be answered directly from the data: how much subrogation actually went by unnoticed last year, or how long attorney letters sat before anyone opened them. Each of those numbers is a count of missed triggers, and together they make the case for orchestration from the operation's own files.

A system that only watches already earns its keep. Nothing important sits unread anymore, and leadership questions stop being answered by sampling and gut feel. It's also the phase today's models are ready for. Reading a claim file and noticing what matters is exactly the kind of work language models have become reliably good at, and the cost of a miss is an alert that didn't fire, which is no worse than the status quo. Every week of watching also builds a record of what the system noticed and what the adjuster then did about it. That record is what justifies the next phase.

Phase Two: Suggest

In the second phase, the system starts making recommendations. The acknowledgment letter shows up drafted. The claim summary is updated and waiting. The reserve review appears as a proposed task with the reasoning attached, and the wage statement request is ready the day a claim turns lost time. So is the status update to the injured worker, who mostly just wants to know that someone is paying attention. Plenty gets corrected before it goes out, and some suggestions get ignored, which is useful information in itself.

Reviewing this output is real work. If the system surfaces 40 suggestions a day to an adjuster who already has 30 diary entries due, nothing has been automated; a backlog has been created with extra steps. The review process has to be staffed and sized like any other part of the operation, and it has to be worth the adjuster's time. The corrections matter most. Every fix an adjuster makes to a suggestion is a lesson, and a system that keeps those lessons gets better every week it runs, while a system that throws them away will be making the same mistakes a year from now. The way out of this phase is consistency, motion by motion: suggestions going through untouched often enough that the approval becomes a formality.

Phase Three: Act Where Possible

The last phase is the one vendor demos lead with, and the one that should come last. Some motions start running on their own, but only the ones with a record behind them, and that record is task by task. Sending an acknowledgment letter unsupervised is a very different proposition from touching a reserve unsupervised, and phase two produces the evidence for that call: how often adjusters approved each kind of suggestion unchanged, and where the overrides cluster. Settlement authority for a new adjuster grows in the same way, with the track record. Autonomy should expand at the pace the record justifies.

The AI models will keep getting better. That doesn't change the design. It just changes how quickly a claims operation moves through the phases.

Where Insurance AI Deployments Stall

Most P&C carriers have AI in production, but almost none achieved adoption because models sit beside claims decisions, not inside them.

Insurance

The insurance industry spent 2024 and 2025 arguing about whether AI worked. That argument is settled, but it produced a strange result.

Datos Insights, reviewing AI implementations across property and casualty carriers this year, puts it plainly: most carriers have AI in production, but almost none have achieved widespread adoption. The top production use cases are all variations on AI reading and summarizing documents. 

The pilot phase is over, but the industry stopped one step later.

In my experience working with VPs of claims at US P&C carriers, this is not a model problem, a budget problem or an ambition problem. It is what happens when AI is installed next to a claim decision instead of inside the process that produces it. The analysis gets better. But the decision takes exactly as long as it always did, because nothing around it moved.

"In Production" Became A Milestone Worth Gaming

Production and adoption are different states, and the industry has quietly started reporting the first while meaning the second.

Production means a model is deployed, serving requests, monitored and owned by someone. It is a real engineering achievement, and it is entirely compatible with nothing changing. A claims organization can have a summarization model in production for 18 months while every adjuster still opens the full file, because the process that governs their work was never rewritten.

Adoption means the work changed. Someone removed a manual assembly step, moved a review earlier in the file, or gave a supervisor a documented basis for an approval that used to rest on recall. The licensed claims professional still makes the determination. What changes is how much of the shift is spent getting to the point where determining anything is possible.

That requires touching process documentation, authority matrices, audit procedures and the core system. Almost none of it is AI work.

Datos Insights identifies data readiness and deployment maturity as the two variables that explain most of the variation in AI outcomes between carriers. I would put the same finding more bluntly. The wall is not in front of the model. It is between the model and everything the carrier already owns.

Summarization Is A Comfortable Place To Stall

The dominant production use case across carriers is reading and summarizing documents. I want to be fair to this: it works, it saves real time, and it was the correct place to start.

It is also structurally safe in a way that should make executives suspicious. Summarization touches no decision authority. It requires no model governance sign-off and no change to the policy administration system. A carrier can run it at scale for two years and never confront the question that actually gates value: is this organization willing to let a machine-produced analysis become part of the documented basis for a determination that a licensed claims professional signs?

That question has an answer, and it is not a technical one. It is a governance answer. Who reviews the analysis, against what standard, with what authority to accept, modify, reject or escalate it, and where that disposition is recorded.

I'd recommend treating a portfolio that is entirely summarization as a warning sign rather than a milestone. If every use case in production is one that no reviewer ever has to put their name to, the organization has not yet tested the thing it needs to test.

The Wall Is the Write-Back Path

Here is the pattern I see most consistently in stalled deployments. The model produces an assessment. The core system requires a transaction. Between those two facts sits a person, retyping.

A claims model produces a severity indicator at first notice of loss. The claims system has no field for it, no way to record which model version produced it, and no place to store what the adjuster did with it. So the indicator is displayed on a screen, the adjuster reads it, applies judgment, and then keys a value the system already recognizes. The analysis is in production. The cycle time is unchanged, because the constraint was never the assessment, and it was never the adjuster's judgment. It was the retyping between them, repeated across every file in the inventory.

This is why integration with the existing core, rather than model accuracy, is what I see stopping these programs. Carriers built AI capability on the read side, where data already flows outward and nothing needs permission. The write side was left for later, and the write side is where governance actually lives.

Six fields decide whether an AI-supported claim file is defensible: the analysis itself, the model version behind it, the evidence and policy language it drew on, its stated confidence and limitations, the reviewer who accepted, modified, rejected or escalated it, and that reviewer's authority level and timestamp. A carrier that cannot store all six has not built decision support. It has built a second screen. Later has arrived, and closing that gap is a core platform project, not an AI project.

McKinsey's research on AI in the insurance industry points at the same structure from the value side. Isolated proofs of concept, the firm argues, tend to lack workflow integration and sustainable benefit capture, while transforming a whole domain can move the bottom line by double digits.

Its budgeting rule is the sentence I'd pin above the desk of anyone sponsoring one of these programs: for every dollar spent developing digital and AI solutions, plan on at least another dollar to reach full adoption and scale. Most carriers I have looked at are running roughly a 10-to-1 ratio in the other direction.

Governance Is The Gate, Not The Brake

There is a common view that regulation is what slows AI down in insurance. I disagree. I think the causality runs the other way.

The NAIC adopted its Model Bulletin on the use of AI systems by insurers in December 2023, and it asks carriers to maintain a documented program covering governance, risk management, internal controls, and third-party AI. State departments have been adopting it since, and more states regulate insurer AI under their own frameworks. None of that prevents a carrier from scaling. What it does is make scope expansion conditional on being able to explain the system.

A carrier that cannot show which model version produced an analysis, on what data, reviewed by whom, and under what authority, is not blocked by the regulator. It is blocked by its own inability to answer the question, and it self-limits to use cases where the question never comes up. That is the summarization trap described from a different angle.

The defense of an AI-supported claim file in litigation is not that the model was correct. It is that the organization ran a controlled process, reviewed the relevant information, documented the reasoning, applied the required authority, and kept ownership of the outcome with a licensed professional. Carriers that can produce that record can widen scope. Carriers that cannot, stay where they are and call it caution.

McKinsey's 2026 survey on AI trust, fielded across roughly 500 organizations, found that nearly two-thirds of respondents cite security and risk concerns as the top barrier to fully scaling agentic AI. It also found that organizations assigning clear ownership for responsible AI averaged a maturity score of 2.6, against 1.8 for those with no clearly accountable function. Governance capability and scaling capability turn out to be the same capability.

What The Carriers That Got Through Do Differently

I don't have inside visibility into every carrier that crossed from production to enterprise deployment, and some of them got there with brute force. But in my experience working with claims and technology leadership on these programs, four patterns recur.

  • They pick a domain, not a use case. Underwriting submission intake end to end, or first notice of loss through assignment. Isolated use cases produce demonstrations. Domains produce a P&L line.
  • They build the write-back path before the model. Analysis, model version, evidence, confidence, reviewer, disposition. If the core system cannot record both what the analysis said and what the human did with it, everything downstream is theater.
  • They fund adoption at parity with build. Process rewrites, authority changes, training, revised audit procedures. This is the unglamorous half, and it is usually the half that is cut when budgets tighten.
  • They instrument the decision, not the model. Model accuracy is an engineering metric. The business metrics are the ones a chief claims officer already reports: cycle time from FNOL to a decision-ready file, the share of files that reach an adjuster fully assembled rather than half assembled, rework and reopen rates, and LAE per claim. Carriers that report the first and none of the rest generally cannot tell whether anything changed.
The Question Worth Asking

The gap between production and adoption will not close by buying better models. It closes when a carrier can point to a determination its people now reach faster, on better-assembled evidence, with the reasoning on the record, and can show a regulator exactly who made it and under what authority.

So here is the question worth putting to your leadership team this quarter. For every AI system you currently run in production: what work was removed from the process it touches, and what does the claim file now show that it did not show before?

If the answer is that a person has better information but performs the same administrative sequence, and the file records nothing about how the analysis was used or by whom, the system is deployed. It has not been adopted, and the difference is where all the value was.

Sources

  1. Datos Insights. "AI Implementations in 2026: The Pilot Phase Is Over for Property and Casualty Carriers." 2026. https://datos-insights.com/reports/ai-implementations-in-2026-the-pilot-phase-is-over-for-property-and-casualty-carriers-ins-2026-102369/
  2. McKinsey & Company. "The future of AI in the insurance industry." https://www.mckinsey.com/industries/financial-services/our-insights/the-future-of-ai-in-the-insurance-industry
  3. McKinsey & Company. "State of AI trust in 2026: Shifting to the agentic era." 2026. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/tech-forward/state-of-ai-trust-in-2026-shifting-to-the-agentic-era
  4. National Association of Insurance Commissioners. "Artificial Intelligence." Insurance Topics. https://content.naic.org/insurance-topics/artificial-intelligence

How AI Is Reshaping Workers’ Comp Strategy

As AI-related layoff fears subside, workers' compensation professionals are leveraging technology to streamline claims while monitoring potential mental health implications.

Insurance

With several states now allowing all employees, not just first responders, to file mental health-related workers' compensation claims, industry professionals are pondering a potential new challenge. What happens if an employee becomes so concerned about AI taking their job that it has an extraordinary effect on their mental well-being?

While it's a fascinating question, the reality is that we haven't seen any evidence of this type of situation occurring in case law thus far. Instead, most workers' compensation professionals are focusing on the positive aspects of AI, using it to streamline the multitude of time-consuming, behind-the-scenes tasks that have slowed down insurers, companies and lawyers for years.

This refocus on the positives of human-supported AI in business combined with recent rollbacks on AI-related layoffs could actually lower stress on employees, as well as reduce the potential for those AI-rooted mental-health related claims.

Job replacement fears allayed

While early employment trends stoked fears of AI replacing human workers, we're now seeing a vastly different scenario play out as companies start to realize that AI, while powerful, has its limitations.

Major employers like Ford and IBM, which conducted AI-related layoffs, rehired employees across engineering, human resources, and customer service this summer to fix quality issues that AI could not address, according to a recent CNBC article. Another report shows that 55% of business leaders who made employees redundant after deploying AI now admit they made the wrong decision.

As companies roll back AI-related layoffs, the renewed emphasis is on creating a workplace where AI and humans co-exist. That reframing is reducing stress for some employees rather than adding to it. Workers are using AI to complete more tasks faster. The end result is that some employees can now spend more time interacting with humans and less time writing emails and staring at computer screens.

Yet while the "AI is taking your job" storyline has subsided in some instances, the case law regarding AI and mental health in the workplace has yet to be written. Workers' compensation boards in most states are very claimant-friendly, and with AI changing so much about how employees do their job, it's within the realm of possibility that we may see stress-induced claims related to AI pop up in the coming months and years.

Benefits Outweigh the Concerns

Despite the negative headlines related to job replacement, AI presents benefits to all parties involved in a workers' compensation claim.

To see how, consider the common denominator in any claim: paperwork. Claimants, claims professionals, and attorneys must comb through thousands of pages of incident reports, medical summaries, specialist notes, imaging studies, surgical records, physical therapy and rehabilitation notes, return-to-work forms, and disability status certifications. Doing so manually typically took months.

AI now expedites the process by reviewing and summarizing reams of digitized files and pulling out the most pertinent information. Lawyers and claims professionals can search for specific terms within those documents to refine their case. These types of efficiencies accelerate the entire litigation process, potentially leading to faster resolution.

AI also benefits injured workers. Some states require workers to show proof of their job search after they've recovered from an incident. With AI, workers can enter their skills, experience, and medical restrictions, then run queries to find jobs that match their unique needs. Carriers and third-party administrators are doing the same, using AI to assist claimants in the job-search process.

For corporate risk managers, AI insights can help develop stronger workplace safety programs. An employer can review its claims database, look for repeat patterns and determine whether injuries are more common at a certain job site, in a specific role or at a certain geographic location. They can then dig into the root causes of incident reports and adjust protocols. If data shows, for instance, that workers are getting hurt due to improper lifting, risk managers can enhance their safe-lifting training program by providing more specific and frequent education, thereby improving worker health and well-being.

AI is also shaping litigation strategy. Attorneys are analyzing historical claims and litigation outcomes with AI to identify trends in settlements, dismissals, and case resolutions. These insights provide additional context that claims professionals and legal teams can use to evaluate their past results and tweak their case strategy going forward.

Human Judgment Remains Vital

While AI can support many steps of the workers' compensation process, it does not replace professional judgment. AI is most effective at analyzing data and identifying patterns. Claims professionals and legal teams should remain in the loop at all times, review all AI outputs for accuracy, and make the final decision whenever data conflicts. The strongest results come from combining AI-driven insights with human expertise.

The fear around AI and job replacement is real, and it may someday lead to a mental health claim from an injured worker. Fears, however, have a way of working themselves out over the long run. We saw this when states first began legalizing medical marijuana. The fear was that it would forever change workers' compensation. The reality was that the legislative changes had a more measured impact.

It's too early to know how much AI will change workers' compensation, but as the fear of job replacement starts to subside, professionals should focus on the positives. When AI increases efficiency, we can give workers what they deserve: a safer workplace and a faster resolution to their claims.

Automation Isn't a Proxy for Progress

Commercial insurance needs better information architecture, not just automation, to support the professional judgment that remains central to underwriting.

Insurance

There is a tendency in insurance technology to use automation as a proxy for progress.

If a process takes fewer human interventions, requires fewer keystrokes or produces a decision more quickly, it is assumed to have become better.

In some areas of insurance, that is a perfectly reasonable objective. Standardized transactions can often be simplified significantly. Routine administration should not consume professional time simply because it always has.

Commercial insurance, however, presents a different challenge.

A significant part of the value in commercial broking and underwriting comes from interpretation: understanding what a business actually does, identifying what matters about the risk, distinguishing routine information from significant detail and applying market knowledge to circumstances that do not always fit neatly into predefined rules.

The opportunity for technology is therefore not simply to remove people from the process. It is to make the information reaching those people materially better.

Automation is not the same as digital improvement

Automation is attractive because its effect is relatively easy to see.

A task that once required five manual steps now requires two. Data moves from one system to another without being re-entered. A document is generated automatically. A workflow progresses without somebody having to send an email.

Those improvements matter.

The difficulty comes when automation becomes the objective rather than one possible consequence of better process design.

If the information entering a workflow is incomplete, inconsistently structured or difficult to interpret, automating the movement of that information does not necessarily solve the underlying problem. It may simply move poor information more efficiently.

Commercial insurance technology should therefore begin one step earlier.

Before asking how much of a process can be automated, it is worth asking whether the information on which that process depends is being collected, organized and retained effectively in the first place.

Commercial insurance remains judgment-intensive

Commercial risks are rarely understood through one variable.

Two businesses with similar turnover and employee numbers can present materially different exposures because of what they manufacture, where they work, the contracts they accept, the territories in which they operate, the property they occupy, the machinery they use or the liabilities arising from their activities.

The same information may also carry different significance depending on insurer appetite, policy class and current market conditions.

That is why experienced brokers and underwriters remain important.

A broker may need to interpret a client's description of an activity and establish what is commercially significant about it. An underwriter may need to understand why an apparently adverse feature exists and whether the wider circumstances alter its relevance.

Client context matters. Market appetite matters. The quality of the dialogue between insurer and intermediary matters.

None of those factors should be treated as evidence that commercial insurance has failed to modernize. They are areas in which professional expertise has economic value.

Technology should help that expertise operate from a stronger information base.

Poor information creates avoidable friction

Many commercial insurance journeys still involve information moving between businesses, brokers and insurers in multiple formats.

There may be proposal forms, spreadsheets, presentations, PDFs, email threads, supporting documents and conversations held at different stages of the process.

There are good reasons for some of this. Different participants require different levels of information, and new questions arise as a risk is examined in greater detail.

But there is also avoidable duplication.

A business may provide basic organizational information several times. A broker may receive important details embedded in correspondence rather than in a consistent proposal structure. Information may need to be manually transferred between systems. A submission may reach an underwriter with significant detail present but with one important point missing or difficult to locate.

Previous information may be technically available but difficult to reuse because it was never stored in a sufficiently structured form.

No single broker, insurer or technology platform experiences all of these problems in the same way. Many organizations already have highly developed processes.

The broader issue is that commercial insurance information is still often organized around individual transactions and systems rather than around the underlying business risk. That creates friction for everybody involved.

Structure before automation

A more useful starting point is structured information.

That does not mean forcing every commercial risk into an inflexible questionnaire. Good structure should actually make a proposal more responsive.

Occupation and industry information can establish context early. Questions can then be presented progressively according to the information already supplied. Data can be captured in defined fields where consistency matters while still allowing explanation where commercial nuance is necessary.

Incomplete areas can be identified more easily. Information already supplied can be updated rather than repeatedly recreated. The resulting proposal can be presented in a consistent format without pretending that every business is the same.

This distinction matters because an online form is not automatically a digital insurance process.

A thirty-page paper proposal converted directly into thirty web pages may technically be digital, but the underlying information design has not necessarily improved.

The stronger question is: what information does the next professional in the process genuinely need, and how can that information be gathered with the least unnecessary burden on the person providing it?

Once that structure exists, automation becomes considerably more useful.

Better information changes the professional conversation

The value of structured information is most apparent when the broker or underwriter receives it.

A broker preparing for a first substantive discussion should ideally be able to identify the nature of the business, its principal activities, obvious exposures, relevant supporting information and any areas that require clarification.

An underwriter receiving a submission should not have to determine whether an important detail exists somewhere in an email chain before considering the risk.

Technology can help establish that starting point. It can organize routine factual information. It can identify missing fields. It can retain information supplied earlier in the process. It can make proposal output more consistent. It can reduce some administrative re-keying.

What it cannot do reliably in every commercial circumstance is decide what every fact means.

That distinction should shape where technology stops.

Where technology should stop

The attraction of end-to-end automation is understandable. A fully digital process can appear cleaner than one containing deliberate human intervention.

But the presence of professional judgement is not necessarily a process defect.

There are points at which somebody should ask another question. There are risks where an experienced broker should challenge how the business has described an activity. There are circumstances where an underwriter should seek additional context rather than allowing a rules engine to produce a simplistic outcome.

There are insurance conversations in which the relationship between the participants contributes materially to the quality of the result.

The objective should therefore be selective automation.

Automate tasks where the result is more reliable, more consistent or less burdensome. Structure information where doing so improves its usefulness. Retain human intervention where interpretation, judgement, advice or negotiation creates value.

Commercial insurance does not need to become less professional in order to become more digital.

Data quality matters more than data volume

There is another risk in digital proposal design: collecting too much.

Because digital systems can store large amounts of information cheaply, there is a temptation to ask every question that might conceivably become relevant.

That is rarely good information design.

A business owner confronted with a long sequence of questions that appear unrelated to the business is not necessarily going to provide better information merely because the system has collected more fields.

Relevance matters.

Questions should be driven, where possible, by context already established about the occupation, industry, activities and exposures involved.

The purpose is not to create the largest possible dataset. It is to create a sufficiently complete and well-structured representation of the risk for the next stage of professional consideration.

Information should be reusable

Commercial insurance information also has a life beyond the initial proposal.

Businesses change, but much of their underlying information persists. Premises, business activities, organizational structures, equipment and other characteristics may remain relevant across different insurance discussions and over successive periods.

Where information can be retained appropriately and kept current, the customer should not always have to start from zero.

The principle is straightforward: confirm and update where necessary rather than repeatedly reconstructing the same business description.

That still requires controls around data accuracy, currency, privacy and purpose. Reuse should not mean assuming that last year's information remains correct.

But the ability to retain structured information provides a much stronger foundation than treating every interaction as an entirely new collection exercise.

One example of the approach

This philosophy is also informing the development of DIBNI, a digital commercial insurance introduction and proposal platform. The work is focused on how commercial information can be structured to support professional review and discussion, rather than attempting to automate professional insurance advice.

DIBNI is only one example of the wider direction. The more important point is that technology should improve the quality and usability of information without displacing the judgement, accountability and relationships that remain central to commercial insurance.

The wider opportunity

Commercial insurance does not need more automation for its own sake. It needs better information architecture around the professional decisions that already matter.

The future model is likely to combine structured technology, professional judgement and commercial relationships. Each plays a different role, and the value comes from making them work together rather than treating one as a substitute for another.

If information is clearer, more consistent and easier to reuse, brokers and underwriters can spend less time reconstructing the basic facts and more time applying expertise where it creates value.

That is a more useful measure of digital progress than simply counting how many human steps have been removed from a process.

Better information enables better professional conversations. For commercial insurance, that may be the most important digital improvement of all. construction.


Gary Huddleston

Profile picture for user GaryHuddleston

Gary Huddleston

Gary Huddleston is the founder of DIBNI and founder and commercial director of Digital Brokers Group Ltd. 

He has more than 25 years' experience across insurance broking, commercial underwriting, business development and executive leadership in the U.K., Ireland and Australia. His current focus includes digital commercial insurance proposal and introduction journeys.

AI Enables Predictive Maintenance

AI-powered predictive maintenance helps manufacturers forecast equipment failures, minimize downtime, and optimize costs.

Insurance

With artificial intelligence continuing to reshape the manufacturing sector and driving its shift toward Industry 4.0, we can see more manufacturers investing in high-impact AI use cases, such as generative design, automated visual inspection, or inventory optimization. More examples of AI in manufacturing include predictive maintenance, a highly popular AI application that can deliver measurable ROI and tangible efficiency gains while reducing the risk of interruptions to the business. 

In short, this use case involves applying AI to analyze how conveyor machines, robotics, vehicles, and other types of manufacturing equipment perform to accurately forecast potential performance issues and address them before they cause malfunctions and breakdowns. This way, manufacturing companies can minimize unplanned downtime, extend machinery lifespans, and optimize maintenance costs.

Gathering and preparing data

To accurately forecast equipment failures, AI-powered predictive maintenance systems need to have a holistic overview of the machine's real-time performance and physical environment, as well as past operational history. This, in turn, requires collecting and merging information from disparate sources, such as real-time insights from IoT sensors and historical data from enterprise systems, such as CMMS, ERP, and MES. As collecting and preparing all this information efficiently is simply impossible with manual methods, modern manufacturers typically delegate this task to AI tools.

For example, manufacturers can deploy AI models directly on local IoT gateways to preprocess raw sensor data instantly at the edge and then transmit it to a centralized cloud storage. Additionally, they can use AI-enabled optical character recognition tools to automatically extract unstructured data from static documents and photos – such as handwritten maintenance logs or scanned and photographed equipment operator notes – standardize it in the format required for the analysis, and also store it in the cloud. 

As manufacturers handle increasing volumes of structured and unstructured information, protecting sensitive data throughout collection, processing, and storage is essential to maintaining security and compliance. A predictive maintenance system can then ingest this historical and real-time data to forecast equipment failures.

Conducting baseline profiling

Even with a rich and high-quality dataset, identifying anomalies or degradation in equipment performance is impossible without a benchmark – often called a baseline profile – that establishes what is "normal" operation for a specific piece of equipment. Instead of creating such profiles manually, manufacturers can program AI algorithms to analyze historical and real-time equipment data, taking into account variables such as equipment workload, age, or speed, to build accurate, tailored baselines for various equipment pieces.

Detecting anomalies and predicting failures

With an established baseline, predictive maintenance systems can continuously collect real-time data streams and compare them against the set parameters. Once an anomaly exceeding the predefined threshold is detected, AI algorithms can automatically classify and map it to known failure patterns, compare it to similar historical examples, and then accurately predict how many operating cycles or days and hours the equipment has left before the breakdown.

Executing diagnostics and root-cause analysis

In addition to predicting equipment failures, AI-powered systems also help manufacturers explore detected issues and identify their root causes. For example, after spotting an anomaly, AI algorithms can compare real-time IoT data with historical data associated with past failures from maintenance logs to pinpoint the specific component and spare part that is starting to wear out. 

AI can also evaluate the exact conditions at the time of the anomaly (e.g., runtime hours, load cycles, or operating temperatures) to determine if a component is wearing out due to operational misuse or mechanical stress. Once the problematic component or root cause of failure is identified, a predictive maintenance system can recommend specific actions – such as load reduction, recalibration, or component replacement – to prevent the issue from recurring.

Automating maintenance scheduling

AI-enabled predictive maintenance systems can automatically generate, prioritize, and optimize maintenance schedules, presenting them for human review and approval, thereby helping manufacturers accelerate administrative maintenance processes. Specifically, after diagnosing a potential issue and calculating time-to-failure, the AI-driven maintenance system can automatically check information related to technician work hours, spare parts inventory, and production line usage, generate a work order, select the optimal time slot for maintenance jobs, and even create step-by-step maintenance instructions for the technicians. 

Subsequently, the generated work order, together with the task priority and a detailed task description, can be automatically sent to the technician's field service management mobile app.

Final thoughts

There are many notable AI use cases in the modern manufacturing industry, and predictive maintenance stands out as one of the most popular and high-impact applications. 

Using AI models to continuously analyze equipment performance data, manufacturers can accurately forecast equipment failures before they occur and address them in a timely manner, therefore reducing unplanned downtime, eliminating routine inspections, and extending the overall usable life of production assets.

Data Centers' Evolving Insurance Needs

Data center investment is set to hit $1 trillion by 2027, but physical constraints and climate risks are reshaping the sector's insurance needs.

Insurance

Annual investment in data centers is projected to double from around US$500bn in 2024 to more than US$1trn as early as 2027. The investment opportunity extends far beyond server halls to electricity generation, grid infrastructure, cooling, networking, and semiconductors.

According to Allianz Research, the US and China are expected to account for around 62% of new global capacity additions through 2030, but the next wave of investment is becoming increasingly global. In Europe, Germany, the UK and Ireland remain major markets, but faster expansion is expected in Spain, Finland and Denmark, where power availability and permitting conditions can be more favorable. Across Asia Pacific, excluding China, installed capacity is projected to increase from around 9GW today to more than 28GW by 2030, with Malaysia expected to grow more than tenfold.

The sector's biggest constraints are increasingly physical rather than financial. Competitive advantage is increasingly determined by access to electricity, grid connections, permitting, specialized equipment and skilled labor. In the US alone, the construction industry faces a shortage of around 439,000 skilled workers, while an estimated 349,000 additional workers may be needed in 2026.

Climate resilience is increasingly a strategic consideration rather than an operational afterthought. Around 79% of global data center capacity is already located in areas exposed to heightened natural catastrophe risk, while 54% is exposed to chronic heat and drought stress. Some of the fastest-growing AI infrastructure markets are also among the most climate-exposed, including Northern Virginia, US, Johor in Malaysia, and Marseille, France. Acute flood, wildfire and wind exposure is highest in the Americas, affecting 86% of capacity, while chronic heat and drought stress is greatest in Asia Pacific, where 89% of capacity is exposed.

Insurance is evolving alongside the sector. As data centers assume a more critical role in infrastructure, comprehensive insurance cover has become a prerequisite for financing many large-scale AI infrastructure projects. Construction costs for a single AI campus can exceed US$20bn, with insured values rising substantially once high-performance computing equipment is installed. The global data center insurance market is projected to grow from around US$11bn today to more than US$24bn by 2030, reflecting rapid capacity expansion, rising insured values and increasing operational complexity.

Demand is expected to extend beyond traditional property cover towards integrated solutions spanning construction, engineering, property, business interruption, cyber and liability, while also creating new opportunities in areas such as energy resilience, operational continuity, and technology risk.

Allianz Commercial analysis of insurance industry data center-related claims shows that fire is the leading driver of loss severity, accounting for well over 50% of around €700mn (US$800mn) worth of losses. Natural catastrophe activity ranks second, followed by willful acts, which include crime and cyber incidents, followed by power failure. Water damage is the most frequent cause of data center claims, followed by willful acts, fire, and equipment breakdown. Business interruption is the primary driver of claims severity by line of insurance, highlighting the significant financial impact of operational downtime.

The data center risk profile is changing as facilities become larger, more complex, and increasingly connected. Hyperscale and colocation of campuses can bring together multiple tenants, construction works, servers, supporting utilities and on-site infrastructure in one physical or operational space. A single event can therefore trigger claims across property, construction, business interruption, liability, cyber, and financial lines. Real-life claims case studies show that in hyperscale facilities, damage to external cooling systems, hot works-related fire damage, and a delay in start-up caused by power disturbances have each resulted in losses in the US$50mn to US$100mn range.

To read the full report, please visit: The data center construction boom: risks and claims trends

Agentic AI Redefines Real-Time Underwriting

Real-time underwriting isn't about instant decisions but leveraging agentic AI and current data to help underwriters assess risk faster.

Insurance

For years, the insurance industry has talked about real-time underwriting as the goal: pull in current data, assess the risk, calculate a price, and deliver a quote almost instantly. But that definition works better for a personal auto policy than it does for a complex multimillion-dollar commercial risk. If insurers want to make real-time underwriting a reality across more of the industry, the definition needs to be broader. The goal doesn't have to be an instant decision, but rather using agentic AI and the most current information available to make better decisions, faster.

That distinction matters. Fully automated, straight-through processing can work when risks are relatively standard. But underwriting doesn't need to be automated from beginning to end to become more real-time. For complex risks, the opportunity is to automate certain areas, bring more current information into the decision, and give underwriters what they need to act faster.

Real-Time Underwriting Looks Different Across the Industry

While real-time underwriting will not look the same across every line of business, the ultimate goal can be the same: accurately assessing risk and acting on that information as quickly as possible. For standardized risks like personal auto, homeowners policies, and small commercial BOP, the result may be agentic AI solutions performing straight-through processing with human oversight. For large commercial risks, where experienced underwriters need to evaluate complex exposures, full automation may not be practical or desirable, but AI tools can expedite a number of manual processes.

Across those different types of risk, three principles can move underwriting closer to real time.

Automate the processes that can be automated. Automation should accelerate the underwriting process without removing the judgment that complex risks require. Agentic AI tools can ingest and structure submission data, identify missing information, check the risk against appetite, and summarize key information. Underwriters then review the recommendations, exceptions, and complex risk factors. By automating the work surrounding the decision, the underwriter can spend more time on the decision itself.

Incorporate more current information into risk decisions. Insurance has traditionally relied heavily on historical information. Loss histories, prior claims, and other historical data remain essential, but the past is not always an adequate predictor of what happens next.

Insurers now have access to an expanding range of external data that can provide a more current view of risk. Telematics, for example, can provide information about actual driving behavior. Climate-related risks can change the relevance of historical property loss data. A property with limited losses over the previous two decades isn't necessarily exposed to the same level of risk today. The challenge is turning that information into something underwriters can actually use by incorporating it into the models and workflows that drive decisions.

Continuously monitor risk after a policy is written. Risk doesn't stand still once a policy is bound. Events can happen during a policy term, including new claims and changes in financial conditions, which can alter the insurers' view of the risk. Rather than waiting until renewal to understand what has changed, insurers can continuously monitor meaningful developments across individual accounts and their broader portfolios.

That doesn't mean every new signal triggers changes in pricing or coverage. Instead, continuous risk assessment gives insurers a more current understanding of the risks already on their books and better information when it is time to make the next underwriting decision, including at renewal.

Putting the Real-Time Mindset Into Practice

None of this requires insurers to rebuild underwriting overnight. Three practical steps can begin moving the organization toward faster, more responsive decision-making.

  1. Incorporate new data into the models that drive decisions. There's an expanding universe of external information. Beyond telematics, connected home devices and geospatial imagery provide more comprehensive information on property risk. Unstructured sources such as news and social media can provide valuable insight into litigation and financial conditions, for larger commercial risks. But more data does not automatically produce better underwriting. Carriers can determine which data meaningfully improves risk assessment and then update rating models, rules engines and underwriting workflows so those signals can influence decisions.

    All of the rating engines don't need to be overhauled at once. Insurers can start with a specific line of business or risk factor, identify an external data source that can improve the assessment of that risk and test how the new information changes pricing and underwriting outcomes. Rating logic should also be flexible enough to consume new data through APIs or other integrations rather than requiring extensive system changes each time a new source is added.

  2. Orchestrate data across the enterprise. Insurers often already possess valuable information that underwriters cannot easily access. Systems such as policy administration, claims, and billing may operate separately, leaving relevant information fragmented across the organization. A significant claim could occur for a commercial account, and an underwriter might not discover it until renewal.

    This is not necessarily a data problem, but an access and orchestration problem. Connecting those workflows can make underwriting more responsive. Insurers should work toward creating a connected data layer that allows information to flow between systems and establishes consistent definitions and formats so it can be used across the organization. APIs and industry data standards can help insurers connect legacy and modern platforms without waiting for a wholesale core replacement.

  3. Automate the work surrounding the decision. Even when the underwriting decision itself requires human judgment, much of the work surrounding it may not. Submission triage is a good example. Rather than having an underwriter immediately begin reviewing every incoming submission, agentic AI powered solutions can first assess whether an account meets basic criteria, identify missing information, and summarize what's been provided. If information is missing, the solution can notify the broker immediately instead of discovering the gap later.

In addition to triaging, agentic AI and automation tools can increasingly handle other manual processes including data ingestion and summarization leaving underwriters to focus on the risks and decisions that actually require their expertise.

Real-time underwriting does not have to mean instant underwriting. It means making underwriting more responsive at every stage, using agentic AI solutions to automate where it makes sense, bringing current information into the decision, and maintaining a more current view of risk after the policy is written. For some risks, that may mean straight through processing. For others, the underwriter is the decisionmaker throughout. But the opportunity across all lines is to give underwriters enhanced information sooner so they can make better decisions.