Rey.id

Rey.id, an AI-first health insurtech, cuts claims from days to minutes with Gemini

Results on Google Cloud
  • Handled 10,000+ daily claims valued ~2M claims worth USD 200M a year at 15% CPU usage using GKE

  • Cut underwriting time from months to minutes with Gemini

  • Found 10% cost savings in pharmacy spend using BigQuery

  • Resolved 93% of support tickets autonomously using Google Cloud

  • Health Claim process from days to under five minutes with Gemini

Rey’s Olvo AI uses Gemini to process 2M annual claims, cutting turnaround to under 5 minutes and automating fraud detection.

Overcoming manual bottlenecks to scale health claims administration

Rey.id began in 2021 as a health subscription provider before expanding into third-party administrator (TPA) services, running claims operations for insurers and employers across Indonesia. Processing health data at scale is often complex: a single claim often arrives as a pages-long bundle of paper documents, hospital invoices, itemized bills, and handwritten doctors’ notes.

Rey.id’s human analysts spent hours manually evaluating disparate provider invoices and checking prices line by line. The company knew it needed a major technological leap, but its initial AI integration attempts stalled. During a pilot run with Indonesia’s state electricity company (PLN) in 2024, Rey.id’s existing architecture achieved only a 40% automated match rate on clinic paperwork. Strict token concurrency and rate limits in their previous cloud provider architecture created major operational bottlenecks, often pushing processing times late into the afternoon and threatening service level agreements.

To build an AI solution that can untangle complex, unstructured medical data without breaking under high volume, Rey.id migrated its core workloads to Google Cloud.

Claims are where insurers win or lose customer trust. We needed AI that reads a 100-page hospital bundle like an experienced adjudicator, at the volume Indonesia’s largest payers run, without patient data ever leaving the country.

Evan Wijaya Tanotogono

Co-founder and CEO, Rey

A dashboard interface from Google Cloud Connect highlighting a presentation for OLVO

Processing complex medical bundles and automating judgment with Gemini

Migrating to Google Cloud enabled Rey.id to build OLVO.ai, a generative AI claims engine. By utilizing Gemini, Rey.id eliminated previous token limitations. Crucially, its massive context window allows the platform to ingest 100-plus-page medical document bundles in a single pass.

With Gemini, a 100-page bundle keeps its thread. It maps a handwritten note on a random page to a diagnosis seamlessly. Google Cloud’s Jakarta data center also ensures we meet strict national data compliance. All of it stays in Jakarta inside a dedicated VPC, encrypted in transit and at rest, with role-based access, a full audit trail, and 24/7 monitoring. Security by design, not as an add-on; it’s how OLVO.ai was designed.

Bobby Siagian

Co-founder and CTO, Rey.id

With Gemini 3.1 Pro, the system executes complex multimodal reading. For example, it seamlessly maps a handwritten drug prescription on page 40 directly to a typed diagnosis on page three. The model structures this unstructured information, correctly maps it to medical codes like ICD-10, and checks line items against provider price lists. To support this scale, Rey.id utilizes Google Kubernetes Engine (GKE) and Pub/Sub to process up to 10,000 claims daily, valued at ~2M claims worth USD 200M in claim value a year for insurers, keeping CPU utilization at a lean 15% and SLA performance at 99.95%.

BigQuery powers the backend data warehouse alongside Knowledge Catalog (formerly Dataplex) and Cloud Storage to analyze drug usage and detect fraud, waste, and abuse (FWA). Operating in a regulated market requires robust security. Running workloads within the Jakarta region using a secure Virtual Private Cloud (VPC) ensures the company complies with ISO 27001 and Indonesia’s Personal Data Protection (PDP) law.

The shift radically transformed the company’s operations. Administrative workload dropped by 30%, allowing analysts to focus on high-level judgment. Turnaround times accelerated dramatically, shrinking a partner’s reinsurance underwriting processing from six months to under 24 hours and claim processing from days to under 5 minutes. At the same time, AI chatbots resolve 93% of customer support tickets autonomously. In analyzing pharmacy pricing variations, Rey.id also identified potential cost reductions of over 10%, proving that enterprise AI can handle both incredible complexity and stringent data compliance.

Rey.id team

Rey.id is an AI-first health insurtech in Indonesia. Its OLVO AI engine automates claims adjudication, fraud detection, and underwriting for insurers and national health payers.

Industry: Healthcare and Insurance

Location: Indonesia

Products: Gemini, Gemini 3.1 Pro, Google Kubernetes Engine (GKE), BigQuery, Knowledge Catalog (formerly Dataplex), Cloud Storage, Virtual Private Cloud (VPC)

Google Cloud