Forty point seven days. That is what J.D. Power measured as the average cycle time from first notice of loss (FNOL) to final payment for U.S. property claims in 2026 - the longest reading since J.D. Power began tracking the metric in 2008 [1]. End-to-end claims processing was supposed to compress that number, not stretch it.
In my experience working with VPs of Claims at mid-to-large U.S. P&C carriers, the gap between what end-to-end claims processing promises on a vendor slide and what carriers actually run in production keeps widening. The technology has improved. The workflows underneath have not.
This article covers what end-to-end claims processing actually means in 2026, where claims workflows break, how AI is changing the handoffs between intake and settlement, and what realistic cycle time looks like by line of business. For the full pipeline view including vendor selection and ROI, see our complete 2026 guide to AI claims processing.
The state of claims processing operations in US P&C (2026)
The U.S. P&C industry wrote $1.06 trillion in direct premiums in 2024 according to NAIC and S&P Global Market Intelligence data [2][3]. Roughly a fifth of that volume passes through claims operations every year - call it $200 billion in losses, expenses, and recoveries flowing through systems that, in many carriers, still rely on three or four manual handoffs between intake and payment.
End-to-end claims processing is the design goal for that pipeline. The idea is straightforward: a claim arrives, gets routed, gets investigated, gets decided, gets paid, and gets recorded - without losing context, ownership, or accountability anywhere along the way. The reality is that most U.S. carriers run versions of this workflow where every transition between stages drops information, requires rework, or both.
Cycle time benchmarks are not moving
The J.D. Power 2026 U.S. Property Claims Satisfaction Study reported 40.7 days average FNOL-to-payment and 29.6 days for repair cycle time [1]. Both numbers are among the longest since tracking began. Auto claims, simpler on average, run faster - typical personal auto cycle time sits in the 12-18 day range for non-total losses, though carriers with manual workflows on commercial auto routinely run 30-60 days.
What hasn’t moved is straight-through processing (STP) rate. Datos Insights (formerly Aite-Novarica) reported industry STP under 10% on claims, with nearly 60% of insurers running no STP at all [4]. The carriers I worked with who broke through that ceiling didn’t buy one product - they redesigned the handoffs.
The four-part claims workflow
Every claim, regardless of line of business, moves through four phases: intake (FNOL), investigation (coverage plus facts plus reserves), decision (settle, deny, partial), and execution (payment, recovery, recording). A connected end-to-end workflow means all four phases share the same data model, the same workflow state, and the same audit trail. Carriers that achieve this report cycle time improvements in the 20-40% range without changing headcount. Carriers that don’t tend to add adjusters every year and wonder why throughput barely moves.
The three handoffs that break claims workflows between FNOL and payout
In my experience working with VPs of Claims, the failure mode in end-to-end claims processing is almost always at a handoff between two systems or two teams. There are three handoffs that cause most of the operational damage in claims portfolios I have audited.
Handoff 1 - FNOL intake to claim file
The first break is between intake and handling. A customer reports a loss via portal, phone, mobile app, or agent. The CSR or intake system captures core facts. Then the adjuster opens the file and starts over. Photos sit in the email thread. The policyholder’s first description is on a recording. The agent’s note is in a CRM. The adjuster rebuilds a claim file from fragments scattered across three systems.
I worked with a Northeast specialty P&C carrier where intake-to-assignment alone added two business days to every claim, even before complexity entered the picture. The intake screen captured loss type and policy ID. Everything else - vehicle damage description, witness contact, prior coverage events - sat in attachments that the adjuster had to find, open, and re-key. After we redesigned the handoff so the claim file inherited the intake artifacts directly, the same FNOL workflow stopped requiring duplicate data entry and adjusters started the actual investigation on day one instead of day three.
Handoff 2 - Investigation artifacts to coverage decision
The second break is between investigation and decision. Adjusters gather photos, statements, repair estimates, police reports, and medical records during investigation. These artifacts live in document management systems, email folders, vendor portals, and sometimes physical files. By the time the adjuster makes a coverage call, half the relevant evidence is in a system the decision log doesn’t reference.
This is where claims leakage shows up. The adjuster doesn’t deny a claim element they should have denied - not because they made a bad call, but because they didn’t have the artifact in front of them when they cleared the file. A connected claim file means the decision screen pulls every artifact into one view, with the policy language that applies highlighted next to the evidence.
Handoff 3 - Settlement to system of record
The third break is between settlement and the final claim record. The payment goes out through a payment processor or treasury system. The reserve closes in the claims system. The recovery, if any, tracks in a subrogation tool. The financial close runs on its own cycle. Audit trails fragment. When a state DOI examiner asks “show me the decision trail on these 50 claims,” it takes three days to assemble what should be a one-click report.
The pattern in carriers that fix this is the same: a single claim file object that maintains continuity of data, timeline, and ownership from FNOL through final accounting close.
How AI changes each handoff in the claims workflow
AI does not fix claims operations by replacing adjusters. It fixes the handoffs by removing the rework adjusters do at each transition. Three capabilities matter in 2026.
Multimodal extraction at FNOL
Multimodal large language models read photos, PDFs, emails with attachments, and even handwritten forms. What used to take an adjuster 4-6 minutes per FNOL - opening attachments, extracting damage details, validating policy applicability - now takes 15-30 seconds in carriers running production multimodal extraction. The adjuster opens a claim file that is already structured, validated against the policy, and ready for triage. The Decerto Claims AI platform handles document and image extraction at this step and feeds structured data into the adjuster workspace before the file is opened.
Fraud signals and reserve recommendation pre-adjuster
Before the adjuster opens the file, the system runs fraud scoring against external databases like ISO ClaimSearch and NICB plus internal patterns, runs initial coverage verification, and produces a reserve recommendation with a confidence interval. According to Deloitte, AI-driven detection and multimodal analysis could save P&C insurers $80-160 billion in fraudulent claims by 2032 [5]. The reserve recommendation is the second-order benefit - a reserve placed at FNOL with structured reasoning shortens reserve adequacy cycles and reduces adverse development. Decerto’s Anti-fraud Solution integrates these signals at intake rather than as a post-payment audit.
Settlement orchestration with payment integration
The settlement handoff is where many carriers still lose cycle time. A simple total-loss auto claim with coverage confirmed and reserves set can sit waiting on a manual payment authorization for three to seven days. A connected workflow with AI-driven decision support shifts this to a rules-based orchestration: approved claims under a threshold pay automatically, claims above the threshold route to a human approver with the full file already assembled. The Higson rules engine handles the threshold logic at sub-millisecond execution speed and gives claims operations the audit-ready rule history that NAIC examiners look for.
Cycle time benchmarks by line of business
Realistic cycle time targets vary substantially by line of business. The numbers below come from J.D. Power data, industry research, and my observations across carrier deployments. Treat them as ranges, not commitments.
Personal auto
Non-total-loss personal auto claims should resolve in 8-15 days at a mature carrier with digital FNOL and AI triage. Total losses run longer - 14-25 days - driven mostly by salvage valuation and title processes that sit outside the claims system. Carriers in the top quartile run closer to 5-10 days on simple physical damage claims with photo-based estimating.
Property
Property claims average 25-40 days in industry data, with J.D. Power’s 2026 reading at 40.7 days FNOL-to-payment for U.S. property claims [1]. The repair cycle adds time even after the carrier-side decision is made. Carriers running connected FNOL-to-vendor-network workflows compress this to 20-30 days. The cycle time bottleneck on property is rarely the claims decision itself - it is the gap between the decision and the contractor schedule.
Commercial and specialty
Commercial lines run 45-90 days on average and longer on complex losses. The end-to-end design helps less on commercial because the investigation phase dominates the cycle. The handoff redesign still matters - but the realistic gain is 10-20%, not the 30-40% seen in simpler personal lines. For deeper context on what defines a truly connected system, see what makes a claims management system truly end-to-end.
NAIC AI Bulletin compliance for claims operations
If your claims operation runs any AI at all - triage, fraud scoring, reserve recommendation, document extraction - you need a documented AI Systems (AIS) Program under the NAIC AI Model Bulletin. The Bulletin was adopted by NAIC on December 4, 2023 and operationalizes the NAIC AI Principles around Fairness, Accountability, Compliance, Transparency, and Security [6]. As of August 2025, 24 U.S. jurisdictions had adopted the Bulletin or a substantially similar approach [7]. Colorado, New York, and California layer their own requirements on top.
What this means for VPs of Claims in practice: every AI-driven decision in your claims pipeline needs documentation of model purpose, training data, validation testing, and ongoing monitoring. Audit trail is not optional. The carriers I worked with who implemented AI without AIS Program governance have spent the last 18 months reverse-engineering documentation they should have built from day one. Don’t repeat that pattern. The Decerto AI for Insurance framework is built to satisfy NAIC Bulletin requirements from the first deployment, not retrofitted later.
CAT readiness - designing FNOL for volume spikes
A connected claims workflow has to survive a catastrophe (CAT) event. Hurricane season, severe convective storms, wildfire - any of these can multiply daily FNOL volume by 10-50x for a week or longer. The carriers that handle CAT well do three things in their FNOL architecture.
First, the FNOL intake auto-scales. Cloud-native intake with stateless ingestion handles the spike without manual provisioning. Second, the triage logic has explicit CAT mode - claims tagged with the CAT event code route to dedicated CAT teams and external adjuster networks instead of the standing queue. Third, the decision thresholds for fast-track payments are pre-configured for the event type. Simple property damage under a defined threshold pays on photo evidence and a signed attestation.
In my experience working with VPs of Claims in coastal states, the carriers that get this right test their CAT FNOL configuration before hurricane season starts. The carriers that get it wrong test it when the first event hits, by which point the operational damage is already done. For the broader operational playbook, see claims lifecycle management best practices for insurers.
FAQ - End-to-end claims processing in P&C insurance
What is end-to-end claims processing in P&C insurance?
End-to-end claims processing is the design and operational model that moves an insurance claim from first notice of loss through final payment without losing context, ownership, or accountability between stages. It covers four phases: intake (FNOL), investigation, decision, and execution. The goal is a single claim file that carries the same data and audit trail from start to close.
How long does it take to process an insurance claim from FNOL to payout?
Average cycle time depends on line of business. J.D. Power reported the U.S. property claims average at 40.7 days FNOL-to-payment in 2026. Personal auto runs 8-15 days for non-total losses at carriers with digital intake. Commercial lines typically run 45-90 days. Top-quartile carriers with connected end-to-end claims processing run 20-40% faster than industry averages on simpler personal lines.
What is the difference between claims automation and end-to-end claims processing?
Claims automation is a technology that removes manual effort from specific tasks like document extraction or routing. End-to-end claims processing is the operating model that ensures all phases of the claim are connected. Automation supports end-to-end processing when it strengthens continuity. Automation alone, without workflow redesign, often just speeds up handoffs that should not exist.
How does AI improve claims processing cycle time for US carriers?
AI compresses cycle time at three handoffs: it extracts data from multimodal FNOL inputs like photos, PDFs, and recordings at intake, it scores fraud and recommends reserves before the adjuster opens the file, and it orchestrates settlement decisions through rules-based approval. Carriers with production AI report 20-40% reductions in simple-claim cycle time after 12-18 months of deployment.
What does NAIC AI Bulletin compliance mean for claims operations?
The NAIC AI Model Bulletin, adopted December 2023, requires insurers to operate a documented AI Systems (AIS) Program covering governance, risk management, and third-party AI oversight. For claims operations running AI on triage, fraud detection, or reserves, this means model purpose documentation, training data records, validation testing, and ongoing performance monitoring. 24 jurisdictions had adopted the Bulletin or similar standards as of August 2025.
Talk to Decerto - 30-minute Claims AI assessment
Every month your claims pipeline runs with broken handoffs is a measurable cost: cycle time you can’t compress, LAE you can’t reduce, adjuster capacity you have to backfill with overtime or contracts. J.D. Power’s 2026 numbers show the operational gap between top-quartile carriers and the rest is widening, not closing.
I run a 30-minute Claims AI operational assessment with VPs of Claims and Heads of Claims at U.S. P&C carriers. It is vendor-neutral, NDA-protected, and you get a written assessment whether or not we ever work together. The first call is technical Q&A with me and a senior architect from the Decerto Claims AI team - your portfolio data, your handoff gaps, your realistic cycle time targets.
Not a sales pitch. No demo loop. Calendar link directly, no form.
Book a 30-minute Claims AI assessment with Marcin Nowak - vendor-neutral, technical, your data.
Sources and citations
- J.D. Power. “2026 U.S. Property Claims Satisfaction Study.”
- National Association of Insurance Commissioners (NAIC). “2024 U.S. P&C Industry Annual Statement Data.”
- S&P Global Market Intelligence. “U.S. P&C Insurance Industry: 2024 Market Review.”
- Datos Insights (formerly Aite-Novarica Group). “Straight-Through Processing in Underwriting and Claims: 2023 Update.”
- Deloitte Center for Financial Services. “AI-Driven Fraud Detection and the P&C Insurance Industry.”
- NAIC. “Model Bulletin: Use of Artificial Intelligence Systems by Insurers.” Adopted December 4, 2023.
- NAIC. “Implementation of NAIC Model Bulletin: Use of Artificial Intelligence Systems by Insurers.”
Subscribe to receive the latest blog posts to your inbox every week.
.png)
.png)
.png)
