An Insurance Carrier Partnered With Notch to Automate FNOL Voice Intake, turning high-volume claims calls into structured, compliant, and faster FNOL workflows

A US insurance carrier was looking to modernize one of the most operationally expensive and sensitive parts of the claims journey: First Notice of Loss intake. The carrier handled a high volume of FNOL calls, many of which followed repeatable patterns but still required human agents to collect information, verify details, document the claim, route the file, and trigger next steps.

The challenge was not simply answering calls faster. FNOL is a regulated, high-stakes workflow. The carrier needed a way to reduce manual work without introducing claims leakage, compliance risk, incomplete intake, unauthorized access, or unsupported coverage statements. Every call needed to be handled consistently, documented accurately, and routed through the right claims path.

The carrier partnered with Notch to automate FNOL voice intake using agentic AI built specifically for insurance operations.

The challenge: long calls, manual intake, and downstream claims rework

Before Notch, FNOL calls were handled primarily by human agents across internal teams and outsourced support operations. A typical claim call could take 15 to 20 minutes, depending on the loss type, policyholder details, coverage verification needs, documentation requirements, and follow-up questions.

This created several operational challenges.

First, intake quality varied from call to call. Even experienced agents could miss details, capture incomplete information, or document facts inconsistently across claim types and jurisdictions.

Second, claims teams often had to rework incomplete intakes. Adjusters spent time clarifying loss details, verifying missing information, reviewing call notes, or correcting structured fields before meaningful claim handling could begin.

Third, the carrier relied on separate transcription and translation vendors, which created additional cost, operational complexity, and data handling overhead.

Fourth, high call volume created labor dependency. Whether handled in-house or through a BPO, every FNOL call required billable human time, even when the interaction was straightforward and could be resolved autonomously.

Finally, the carrier needed strong controls around privacy, compliance, and decisioning. The FNOL process involved sensitive personal information, jurisdiction-specific handling requirements, coverage-related questions, and claim routing rules. Any automation needed to be auditable, defensible, and safe by design.

The solution: AI agents that take FNOL calls from start to structured claim handoff

Notch deployed AI voice agents to handle FNOL intake calls directly with policyholders, claimants, brokers, and other authorized parties. The agents were designed to collect loss information, ask dynamic follow-up questions, verify required details, identify missing information, and create a structured intake record that could move into the carrier’s claims workflow.

Instead of treating the call as a simple transcript, Notch converted the interaction into operational data. The AI agent captured claimant details, policyholder information, loss facts, date and time of loss, location, involved parties, injury or damage indicators, supporting context, and next-step requirements.

The agent was governed by deterministic business rules, jurisdiction-specific requirements, and carrier-defined claims workflows. It did not make unsupported coverage or claim decisions. When a call required human judgment, fell outside approved rules, triggered a high-risk condition, or involved an unauthorized user, the interaction was escalated through a logged path.

Every interaction was orchestrated by ADAM, Notch’s operating layer. ADAM reads interactions, identifies escalation patterns, and helps build new agents or workflow improvements to close gaps over time. As more FNOL calls are handled, the automated footprint compounds and exception volume shrinks.

How the FNOL workflow worked

A policyholder or claimant called to report a loss.

The Notch voice agent authenticated the caller where required, confirmed the purpose of the call, and began structured FNOL intake.

The agent collected the required claim details through a natural conversation, adapting follow-up questions based on line of business, loss type, jurisdiction, and carrier rules.

Information was transcribed, translated when needed, and converted into structured claim fields.

The agent verified required information, flagged gaps, and avoided unsupported coverage or claim outcome statements.

High-risk or out-of-scope cases were escalated to a human team with the call context, transcript, structured data, and reason for escalation.

Completed FNOL records were passed into the carrier’s claims systems, giving adjusters a cleaner starting point and reducing downstream rework.

The outcome: faster FNOL, lower operating cost, and stronger claims readiness

With Notch, the carrier was able to automate a significant portion of FNOL voice intake while maintaining the controls required for claims operations.

Across production deployments, Notch has achieved 70% to 73% average autonomous resolution of customer interactions, meaning the majority of eligible interactions are completed without human intervention. For FNOL, this directly reduces dependency on in-house or BPO labor by offsetting call volume that previously required human handling.

The carrier also benefited from faster intake. Compared with a fully human baseline, Notch has delivered a 6x faster median time-to-resolution across production deployments. For a traditional 15 to 20 minute FNOL call, that means a significantly shorter interaction, giving policyholders and claimants their time back while accelerating the start of the claims process.

The business case was built across several measurable levers: autonomous FNOL completions, direct BPO hour offset, reduced transcription and translation vendor costs, fewer adjuster rework cycles from incomplete intakes, and reduced average handle time for calls that still required a human agent. Across production deployments, Notch has delivered 200% ROI within 12 months, with full payback typically occurring between months 4 and 8 depending on BPO rate and volume in scope.

The carrier also consolidated transcription and translation costs. Because transcription and translation are native to Notch, carriers using third-party vendors for these services have been able to eliminate separate contracts after deployment.

The deployment was designed for enterprise insurance requirements, including VPC deployment, PII safety, field-level RBAC, data-zone selection, retention and deletion controls, and auditable decisioning. In similar VPC deployment environments, Notch has gone live in as little as 7 weeks.

Built for claims safety, privacy, and defensibility

Notch is designed to reduce leakage across the full FNOL decision path. In practice, this means preventing AI agents from making unsupported coverage or claim statements, stopping unauthorized users from accessing or progressing claim workflows, blocking high-risk actions unless deterministic business rules are satisfied, enforcing jurisdiction-specific handling requirements, and ensuring claims move through structured intake, correct routing, and logged escalation paths.

The platform is built with PII safety in mind. Notch supports data-zone selection, including all-US deployment, automated data deletion and retention rules, API-based deletion, and vendor controls designed to support GDPR and other applicable data requirements.

For carriers using Guidewire or similar claim systems, Notch supports field-level RBAC so that sensitive PII fields are visible only to roles with explicit access. Audio recordings and transcripts can follow the same retention and deletion rules as other claim documentation.

If payment information is collected during claims calls, such as deductible collection, Notch supports pause-resume recording controls and dedicated PCI-compliant flows.

Core use case: FNOL Voice Intake Automation

Notch automates FNOL voice calls by turning natural conversations into structured, compliant claim intake records. The AI agent collects required information, adapts questions based on loss type and jurisdiction, identifies missing details, routes exceptions, and hands off completed claim records to the carrier’s downstream systems.

The result is a faster, more consistent FNOL process that reduces manual call handling, lowers BPO and vendor costs, improves intake completeness, and gives claims teams cleaner files from the first touch.

Bottom line

By automating FNOL voice intake with Notch, the carrier moved from a labor-heavy, manually documented claims intake process to an AI-powered workflow that can resolve the majority of eligible interactions autonomously, shorten call duration, reduce vendor and BPO costs, and improve the quality of information entering the claims process. With ADAM orchestrating every interaction, the carrier gained an operating layer that continuously learns from escalations, expands automation coverage, and keeps every decision auditable, defensible, and aligned with insurance-specific rules.

‍