Why AI Matters for Insurance

The hidden cost of manual underwriting and claims processes

The cost of manual document processing in insurance is not just measured in hours — it shows up in your combined ratio, your loss ratio, and your submission-to-bind rate.

Underwriting throughput is constrained. A commercial lines underwriter can manually review 3 to 5 new submissions per day. Each submission requires reading the ACORD application, loss runs, supplemental questionnaires, and broker commentary, then keying risk data into the policy administration system. The result: carriers routinely decline or ignore 60-80% of submissions simply because underwriters lack time to review them. Those are potential premiums walking out the door.

Claims leakage is staggering. The insurance industry loses an estimated $30 billion or more annually in the US alone to claims leakage — overpayments, missed subrogation opportunities, undetected fraud, inaccurate reserves, and settlement inefficiencies. McKinsey estimates that 5-10% of every claims dollar is leakage. On a $50 billion book of business, that is $2.5 to $5 billion in preventable loss.

Talent is trapped in data entry. Senior underwriters who should be using their judgment to price complex risks are instead spending their days copying data from PDFs into rating systems. Experienced adjusters who should be investigating complex claims are drowning in routine FNOL documentation. Your most expensive people are doing your least valuable work.

Legacy systems are the silent bottleneck

Most carriers operate on policy administration systems that were implemented 10 to 20 years ago. Guidewire, Duck Creek, Majesco, and legacy mainframe systems process billions in premium, but they were not designed for the AI era.

Data is trapped in silos. Underwriting data lives in one system, claims data in another, billing in a third, and reinsurance in a fourth. Getting a unified view of a single insured across all their policies and claims requires manual querying or report-writing across multiple platforms.

Integration is painful. Many carriers still rely on flat file extracts, batch processing, and manual rekeying to move data between systems. A submission that arrives as an email attachment from a broker must be manually entered into the underwriting workbench. A claim that is reported by phone must be manually documented in the claims management system.

The technology debt is enormous. Carriers know their systems need modernisation, but a full policy administration system replacement is a multi-year, multi-hundred-million-dollar project. Most carriers cannot wait for a system replacement to start benefiting from AI. The good news is that they do not have to — AI can sit on top of existing systems, processing documents and generating outputs that humans then enter into legacy platforms.

The competitive landscape — InsurTechs are not waiting

AI adoption in insurance is no longer experimental. The carriers and InsurTechs that have moved first are setting new benchmarks that traditional carriers will be measured against.

Lemonade built its entire operating model around AI. Their AI Jim chatbot handles claims from FNOL to payment in as little as three seconds for straightforward renters and pet insurance claims. Their combined ratio is still challenged, but their per-claim operating cost is a fraction of traditional carriers.

Root Insurance uses AI and telematics data to underwrite auto insurance based on actual driving behaviour rather than traditional rating factors. Their loss ratio performance on their best-segmented risks demonstrates what granular AI-driven underwriting selection can achieve.

Hippo combines AI with IoT sensor data for homeowners insurance, using smart home devices to prevent claims before they happen and using AI to streamline the underwriting process from application to bind.

Coalition applies AI to cyber insurance underwriting, continuously scanning their policyholders' digital attack surfaces and adjusting risk assessments in real time — something that would be impossible with traditional underwriting approaches.

For traditional carriers, the threat is clear. These InsurTechs are not just faster — they are operating at fundamentally different cost structures. When a Lemonade processes a claim at one-tenth the cost, the combined ratio pressure on traditional carriers intensifies. The response cannot be "work harder" — it has to be "work smarter with AI."

The combined ratio math that makes AI urgent

The combined ratio is the metric that defines carrier profitability, and it is under sustained pressure from every direction.

Loss ratios are climbing. Social inflation, nuclear verdicts, climate-driven catastrophe frequency, and medical cost inflation are all pushing loss ratios higher across most lines of business. The P&C industry average loss ratio has been above 65% in recent years, with many carriers running above 70%.

Expense ratios remain stubbornly high. Despite decades of technology investment, the industry average expense ratio hovers around 27-30%. Manual underwriting workflows, paper-based claims handling, and duplicative compliance processes all contribute. Every point of expense ratio improvement drops directly to the bottom line.

AI directly attacks the expense ratio. If AI can reduce underwriting processing time by 40%, improve claims adjuster productivity by 30%, and automate 50% of routine compliance document review, the impact on the expense ratio is measurable in points, not basis points. On a $1 billion premium book, a single point of combined ratio improvement is $10 million.

And AI can improve the loss ratio too. Better underwriting selection (through more thorough submission review), faster fraud detection, improved subrogation recovery, and more accurate reserving all contribute to loss ratio improvement. The carriers that deploy AI effectively will improve both sides of the combined ratio simultaneously.

The insurance talent gap is widening

Insurance faces a workforce crisis that makes AI adoption not just beneficial but necessary.

The great retirement wave. The average age of an insurance professional in the US is over 40, and a significant percentage of the experienced workforce — senior underwriters, claims managers, and actuaries — will retire within the next decade. The industry needs to replace decades of accumulated expertise.

Young talent is scarce. Insurance struggles to compete with technology companies and other industries for top talent. The perception of insurance as slow-moving and paper-heavy is a recruiting liability. Carriers that can offer modern, AI-augmented workflows have a significant advantage in attracting younger professionals.

AI bridges the experience gap. When a junior underwriter can use AI to quickly analyse loss history, compare a submission against the carrier's appetite guidelines, and surface similar risks from the historical book, they can produce work product that approaches the quality of someone with 15 years of experience. AI does not replace judgment — but it compresses the learning curve dramatically.

The professionals who combine insurance expertise with AI fluency will be the most valuable people in the industry. An underwriter who can direct AI to triage a stack of submissions in minutes. A claims adjuster who uses AI to identify subrogation opportunities across their caseload. An actuary who uses AI to review and summarise competitor rate filings. These are the people carriers will fight to retain.

Key takeaways

1. Insurance is drowning in documents — ACORD forms, submissions, loss runs, claims files, bordereaux, policy forms — and most processing is still manual.
2. The cost is measured in combined ratio points — manual processes drive claims leakage ($30B+ annually), underwriting throughput constraints, and high expense ratios.
3. Legacy systems constrain but do not prevent AI adoption — AI can sit on top of existing policy administration systems and deliver value immediately.
4. InsurTechs are resetting cost benchmarks — Lemonade, Root, Hippo, and Coalition operate at fundamentally different cost structures enabled by AI.
5. The talent gap makes AI essential — with experienced professionals retiring and young talent scarce, AI is how carriers maintain underwriting quality and claims efficiency.

Next up: How AI Works — Insurance Edition.