What challenges did Frende face before AI implementation?

Needed to support a growing customer base without increasing employee headcount. Customer service team overloaded by hundreds of daily inquiries across life, home, auto, and travel insurance lines. Misrouted emails led to duplicated reviews and slower customer responses. Manual review of non-standard vendor invoices (often handwritten) slowed down payment cycles. Maintaining high customer satisfaction while scaling operations.

What was the impact of the AI solution?

15% of Frende's total work performed by digital workers. Reduced customer dissatisfaction by 20%. Net Promoter Score (NPS) increased to 88. No additional hiring required as the customer base grew. Most vendor invoices are processed and paid within a single day.

Source

Microsoft

Frende Streamlines Claims and Invoice Processing with Digital Workers

Use case typeClaims automationUpdated Jun 13, 2026

Frende, an insurance company in Norway, sought to improve customer service and claims handling efficiency without increasing staff. Facing hundreds of daily customer inquiries and vendor invoices, they experienced issues like misrouted emails, duplicated reviews, and slow invoice processing. By deploying SS&C Blue Prism digital workers integrated with Microsoft AI technologies, Frende automated the triage, summarization, and routing of claims emails and automated non-standard invoice processing from vendors. Digital workers extract information using large language models to interpret unstructured data, summarize claims, and validate invoice details. This solution allowed Frende to handle a growing customer base without hiring more claims handlers, reduced customer dissatisfaction by 20%, enabled vendor payments within a day, and achieved an NPS score of 88. Frende’s approach is now foundational for further automation, including end-to-end travel claims processing. The article outlines the business and technical context, customer and partner involvement, and clear metrics for business impact, including automation share and customer satisfaction.

Organization: Frende
Industry: Insurance
Location: Norway
Published: December 2024

Reported outcomes

−20%

quantified impactOther quantified impact

15%quantified impact

Strategic outcomes

Speed & agilityAutomated claims email triage and routingNew product / capabilityAutomated handwritten invoice processingScale & capacityGrew customer base without hiringCustomer experience & trustImproved customer satisfaction and trust

Primary read

Use case focus

Showing 3 of 4

1Automated Claims Triage and Assignment
2Automated Non-Standard Invoice Processing
3AI-driven Email Summarization and Routing in Insurance

Needed to support a growing customer base without increasing employee headcount.
Customer service team overloaded by hundreds of daily inquiries across life, home, auto, and travel insurance lines.
Misrouted emails led to duplicated reviews and slower customer responses.
Manual review of non-standard vendor invoices (often handwritten) slowed down payment cycles.
Maintaining high customer satisfaction while scaling operations.

Implemented SS&C Blue Prism digital workforce integrated with Microsoft AI.
Digital workers use Machine Learning to categorize emails and extract relevant information.
Large Language Models employed to summarize, interpret, and route claims emails for resolution.
Automated extraction and validation of vendor invoice details (including handwritten invoices) using Generative AI.
Freed employees to focus on complex claims and customer communications.

Technologies

Microsoft AI Azure AI Large Language Models Machine Learning Generative AI

15% of Frende's total work performed by digital workers.
Reduced customer dissatisfaction by 20%.
Net Promoter Score (NPS) increased to 88.
No additional hiring required as the customer base grew.
Most vendor invoices are processed and paid within a single day.

Architecture

Digital workers are trained with a Machine Learning model to extract inquiries and attachments from a central mailbox. Each inquiry is sent to a Large Language Model for summarization, key information extraction, and interpretation. The processed inquiry is then routed to the appropriate claims handler. For vendor invoices, digital workers extract invoices from the database, pass them to the LLM for data interpretation, validate extracted data, and hand off validated invoices to the accounts payable team.

Implementation partners1

Blue Prism

Sources & evidence1

Groundedness: Unavailable

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence