Top european insurer accelerates claims notification with Druid AI agents

Why this matters now

In insurance, claims notification is where customer trust is tested first. It is also where manual processes create the most friction: limited operating hours, repetitive data entry, frequent data-processing errors, incomplete submissions and slow handoffs.

The company set out to modernize this critical touchpoint by moving from reactive support to a structured, always-available, agent-led experience that makes claims initiation faster, simpler and more reliable.

Their main goals were to:

• Streamline the First Notice of Loss (FNOL) process and accelerate claims handling.
• Extract information from documents, reduce manual typing and automate data collection.
• Provide policyholders with 24/7 support to submit claims notifications.
• Validate customer and policy data in real time to avoid downstream delays.

The challenge

The insurance company needed to remove a major operational constraint: claims notification was tied to standard business hours, leaving customers without immediate support when incidents happened at night or over the weekend.

At the same time, manual intake created persistent data-quality issues. Missing details, incorrect entries and back-and-forth clarifications delayed file opening and slowed the path to settlement.

Any solution also had to work inside a complex insurance environment. It needed to connect securely with legacy core systems, validate policies in real time and ensure every notification was both technically sound and compliant.

On top of that, the company had to absorb spikes in volume, especially during severe weather events or periods with a high number of incidents, without overloading human teams.

The solution

The company introduced an AI notification agent designed to streamline claims notification end to end, through the following capabilities:

• OCR and document intelligence
  The AI agent uses OCR to read and extract key data from IDs, vehicle registration documents and other supporting files. Claim forms are pre-filled automatically, reducing effort for the customer and limiting manual-entry errors. The result is a faster, simpler claims notification experience with less administrative burden at a high-stress moment.
• Real-time policy validation
  Through secure API integration with the company’s backend systems, the AI agent verifies policy status and coverage in real time, helping customers proceed with confidence while reducing incomplete or ineligible claims.
• Autonomous claim intake
  The AI agent collects the key information and evidence needed to open a claim, validates completeness and creates the claim file in the insurer’s core system, turning FNOL into a structured digital workflow.
• WhatsApp accessibility
  Customers get the same high-quality experience on WhatsApp, so claims notification is no longer limited to office hours or a single touchpoint.

Learnings

Agentic AI delivers the most value in high-stakes processes where speed matters but accuracy is non-negotiable.

Customers are willing to use AI for more complex tasks when the experience reduces effort rather than adding steps. OCR played an important role here by simplifying document handling and reducing manual input.

Meeting customers where they already are is truly valuable. By making the service available on WhatsApp, the insurance company removed the need for a dedicated app and made support easier to access in real time.

Next, the AI agent gets a voice. Voice capabilities that support natural-language conversations will be added to the website, for a seamless and accessible customer experience.

Key wins

• Intent understanding
  Natural language processing accurately interprets customer intent, even with varied phrasing or incomplete information, ensuring smooth interactions.
• Intelligent input validation
  AI-powered validation ensures data accuracy by checking formats, cross-referencing information and prompting for corrections when needed.
• Dynamic flow adaptation
  Conversation flows adjust in real time based on claim type, customer responses and contextual factors, creating personalized experiences.