We value your privacy
We use cookies to improve your experience, analyze site traffic, and personalize content. By clicking "Accept All" you consent to our use of cookies. You can manage your preferences at any time. Privacy Policy
How a National P&C Carrier Automated 85% of First Notice of Loss, and Turned Claims Into a Competitive Advantage.
2.1 million policyholder interactions. $14.2M in cost avoidance. FNOL completion time cut from 34 minutes to 6.4 minutes. This is the story of how Pypestream helped one of America's largest property and casualty carriers transform its most critical customer touchpoint.
A carrier with a claims experience that didn't match its brand promise.
With over 4.8 million active policies across auto, homeowners, and commercial lines in 38 states, this carrier had built a strong market position on the promise of being "there when it matters most." Its underwriting and pricing teams were industry-leading. Its claims experience was not.
The average FNOL call lasted 34 minutes, required three transfers between departments, and left claimants on hold during the most stressful moments of their lives. CSAT scores for FNOL interactions were 2.3 out of 5, the lowest of any customer touchpoint. Worse, 18% of FNOL records contained data entry errors that delayed claim assignment and increased total claim cost.
The carrier's Chief Claims Officer had a mandate from the board: transform FNOL from a liability into a differentiator. The solution had to be compliant across 38 states, capable of handling CAT surge volumes, and empathetic enough to serve claimants in their worst moments. Pypestream was the only vendor that proposed all three.
Pypestream's AI agent Vera handles 2.1M policyholder interactions annually across 38 states.
Four problems that defined the claims experience.
FNOL as a first impression, and a broken one
First Notice of Loss is the most emotionally charged moment in a policyholder's relationship with their insurer. The carrier's average FNOL call lasted 34 minutes, required three transfers, and left claimants on hold during the most stressful moments of their lives. CSAT scores for FNOL interactions were the lowest in the carrier's entire operation at 2.3 out of 5.
Claims leakage from manual data entry
Every FNOL call resulted in a claims adjuster manually entering incident details into the carrier's core system. Data entry errors were present in 18% of FNOL records, requiring rework that delayed claim assignment by an average of 2.3 days. Each day of delay increased total claim cost by an estimated 4.5% due to compounding repair and rental costs.
Surge capacity during catastrophic events
During CAT events, hurricanes, wildfires, hailstorms, inbound FNOL volume spiked 600–900% within 24 hours. The carrier's contact center had no elastic capacity. During a major Gulf Coast hurricane, average hold times exceeded 3 hours. Policyholders who couldn't get through filed complaints with state insurance commissioners, resulting in regulatory scrutiny.
Compliance across 38 states
The carrier operated in 38 states, each with distinct requirements for claims acknowledgment timelines, disclosure language, and claimant rights notifications. Ensuring every AI interaction met state-specific compliance requirements, while remaining conversational and empathetic, was a technical and legal challenge that had blocked prior AI initiatives.
Vera: A claims AI built for empathy and compliance at scale.
Pypestream designed Vera as a named AI agent with a distinct persona: calm, clear, and efficient, the kind of claims experience that makes a policyholder feel supported rather than processed. Vera's conversation flows were co-designed with the carrier's claims operations team and validated by a panel of actual claimants who had filed claims in the prior 12 months.
The compliance rules engine was the technical centerpiece. Vera dynamically adjusted her disclosure language, notification timelines, and claimant rights statements based on the policyholder's state of residence, ensuring every interaction met state-specific regulatory requirements without requiring a human compliance review. The carrier's regulatory affairs team approved the engine after a six-week review.
The CAT surge protocol transformed the carrier's catastrophe response capability. When a CAT event was declared, Vera automatically activated a simplified FNOL flow designed for high-volume, high-stress conditions. Proactive outbound notifications were sent to all policyholders in affected ZIP codes within 90 minutes of the CAT declaration, before most of them had even assessed their damage.
Vera's compliance rules engine covers all 38 states of operation, with dynamic disclosure language, state-specific acknowledgment timelines, and claimant rights notifications. Every interaction is logged with a full audit trail. Pypestream's SOC 2 Type II certification covers all data processing and storage. The carrier's compliance team conducts quarterly audits, zero regulatory incidents in 24 months of production.
From compliance approval to full deployment in 17 weeks.
Pypestream's compliance team worked with the carrier's legal and regulatory affairs team to map state-specific FNOL requirements across all 38 states of operation. A compliance rules engine was designed to dynamically adjust disclosure language and notification timelines based on the claimant's state of residence.
Pypestream's Client Services team embedded with the carrier's claims operations leadership to map 28 FNOL scenarios across auto, homeowners, and commercial lines. Each scenario was designed with empathy-first conversation flows, validated by the carrier's customer experience team and a panel of actual claimants.
Deep integration with the carrier's Guidewire ClaimCenter was completed, enabling Vera, the carrier's AI agent, to create, update, and assign claims records in real time. Integration with the carrier's preferred vendor network enabled Vera to schedule inspections and rental vehicles within the FNOL conversation.
A dedicated CAT surge protocol was designed: when the carrier's catastrophe management team declared a CAT event, Vera automatically shifted to a high-volume mode with simplified FNOL flows, proactive outbound notifications to policyholders in affected ZIP codes, and real-time status updates for previously filed claims.
Full deployment across web, mobile app, SMS, and IVR. Vera handled 85% of FNOL volume in the first month without human involvement. The claims team was redeployed to complex liability investigations, large-loss coordination, and CAT field operations.
"Claims is the moment of truth for any insurer. We had been failing that moment for years, 34-minute FNOL calls, three transfers, claimants on hold during the worst days of their lives. Vera changed that. Our FNOL CSAT went from 2.3 to 4.8 in six months. Our claims team is now doing the work they were trained to do, complex investigations, large-loss coordination, CAT field response. The routine work is Vera's. And she's better at it than we were."
Outcomes at 12 months post-deployment.
85% of FNOL interactions completed by Vera without human involvement, up from 0% at deployment.
Average FNOL completion time fell from 34 minutes to 6.4 minutes, with data accuracy improving from 82% to 99.1%.
Claims leakage from data entry errors dropped by 94%, saving an estimated $3.8M annually in rework and delay costs.
During a major Gulf Coast CAT event, Vera handled 94% of FNOL volume at peak surge, with zero hold times and a CSAT score of 4.6.
FNOL CSAT scores rose from 2.3 to 4.8 out of 5, making claims the highest-rated customer touchpoint in the carrier's operation.
Zero regulatory incidents across 2.1 million interactions in 24 months of production operation across 38 states.
See AI that doesn't just respond. It resolves.
Peek inside real systems of AI agents running at scale. Learn how AI agents complete customer workflows across systems, with the control, precision and outcomes you actually care about.