Upstage
Enterprise AI platform for document intelligence and LLM-powered workflow automation in insurance and financial services
Overview
Upstage is an enterprise AI solutions provider focused on document processing and large language model deployment for carriers, TPAs, and brokers operating in insurance, finance, and healthcare. Founded in October 2020 by former Naver Clova AI executives, the company builds proprietary LLMs and document intelligence engines that extract structured data from unstructured documents with high accuracy.
The core platform addresses the USD 25.7 billion annual cost of insurance claims adjudication by automating document-heavy workflows. Its Document Parse engine converts PDFs, scanned files, ACORD forms, CMS-1500, UB-04, EOB, and FNOL documents into machine-readable structured text at roughly 0.6 seconds per page. The Information Extract module pulls structured data from invoices, claims submissions, and financial documents to support underwriting and claims automation. Upstage Studio, launched in April 2026, adds a no-code agentic layer for deploying document extraction workflows without engineering resources.
The company has raised USD 157M across six funding rounds, most recently a USD 45M Series B Bridge in August 2025 from Korea Development Bank, Amazon, and AMD. Upstage has delivered over 100 enterprise AI solutions, processes 240,000+ documents daily at Hanwha Life Insurance, and secured a strategic partnership with AWS for APAC and US market expansion. It has been named to the CB Insights AI 100, InsurTech 50, and FinTech 100, and Fast Company World's Most Innovative Companies 2026.
Products & Services
Solar Pro 3
Upstage's flagship large language model optimized for large-scale enterprise deployments requiring strong reasoning and instruction following. Solar Pro 2 (released July 2025) was classified as South Korea's first global frontier AI model, reporting benchmark performance above GPT-4.1 and DeepSeek V3 on comparable tasks. Available via REST API through the Upstage Console and AWS Marketplace.
Key Features
- Optimized for high-volume, low-latency enterprise inference
- On-premises deployment option for data sovereignty requirements
- Available on AWS SageMaker, Trainium, and Inferentia infrastructure
Target Users: Enterprise carriers, financial institutions, and regulated-industry operators requiring sovereign or on-premises AI deployment
Solar Mini
A 10.7 billion parameter small language model designed for fast, cost-efficient inference in document classification and extraction tasks.
Key Features
- Lightweight architecture for high-throughput, cost-sensitive workflows
- Available via Upstage Console API and AWS Marketplace
- Deployable as a component within larger agentic pipelines
Target Users: Mid-market insurers and TPAs processing high document volumes with latency and cost constraints
Document Parse
Converts PDFs, scanned documents, handwritten notes, emails, and specialized insurance forms into machine-readable, LLM-ready structured text.
Key Features
- Processes approximately 0.6 seconds per page (100 pages in under a minute)
- 96%+ accuracy across ACORD forms, CMS-1500, UB-04, EOB, and FNOL documents
- Handles mixed-format input including printed, handwritten, and digitally generated content
Target Users: Carriers, TPAs, and brokers automating underwriting submissions, claims intake, and policy management
Information Extract
Pulls structured data fields from unstructured text in invoices, insurance claims, contracts, and financial documents.
Key Features
- Configurable extraction schemas for specific document types
- Supports claims processing, policy management, and underwriting automation workflows
- Integrates with existing core systems via REST API
Target Users: Claims departments, underwriting teams, and operations groups at carriers and TPAs
Upstage Studio
A no-code agentic AI platform launched April 2026 that deploys AI agents to extract, validate, and deliver structured data from complex documents.
Key Features
- No-code workflow builder for document extraction pipelines
- Agent-based validation layer for data accuracy verification
- Designed for insurance, finance, and legal document workflows
Target Users: Operations and IT teams at carriers and brokers deploying document automation without dedicated ML engineering
Syn Pro
A locally trained LLM co-developed with Karakuri and optimized for the Japanese language market. Certified as a domestic foundation model by Japan's Ministry of Economy, Trade and Industry (METI) in November 2025.
Key Features
- Japanese-language optimized for financial and insurance document processing
- METI-certified domestic foundation model status in Japan
- Designed for Japanese market regulatory and data sovereignty requirements
Target Users: Japanese insurers and financial institutions requiring domestic AI infrastructure
AI Space
An enterprise document search and analysis tool for knowledge management and research workflows.
Key Features
- Semantic search across large enterprise document repositories
- Supports policy, claims, and compliance research use cases
Target Users: Underwriters, compliance teams, and claims adjusters requiring rapid document retrieval
At a Glance
- Founded
- 2020
- Headquarters
- Yongin, South Korea
- Employees
- 201-500
- Funding
- Series B
Category & Focus
- Category
- Document Management
- Subcategories
- AI Document Processing Claims Automation LLM Platform Information Extraction
- Insurance Verticals
- Life & Annuity P&C Commercial Health
- Target Customers
- Carriers, TPAs, Brokers
Customers
- Hanwha Life (South Korea) -- Document Parse for 5M historical claims; 96%+ accuracy, 240,000+ documents daily, 70% reduction in manual processing, 50% lower infrastructure costs
- Samsung Life Insurance -- OCR Pack AI solution for financial document processing
- Beach and Company Insurance -- AI-assisted policy review reducing manual hours per policy
- AIA -- Enterprise AI client (referenced on company homepage)
- KB Kookmin Bank -- Enterprise AI client (referenced on company homepage)
- Verra -- End-to-end AI-driven data extraction
- Sendbird -- Solar LLM integration for Smart Assistant customer service
Similar Companies
-
SybrinHyperautomation and intelligent document processing platform for insurance, financial services, and telecommunications -
EXPERLOGIXCPQ, document automation, and digital commerce software for enterprise organizations. -
SignaturitDigital signature and transaction management platform for the insurance industry -
SSideDrawerSecure digital vault and document collaboration platform for financial services firms and wealth advisors
-
SSense6AI-powered document extraction and explainable decision-making for regulated industries
-
RETJAAI-powered insurance policy analysis for brokers and carriers
Last updated: 2026-06-19