Retail SaaS / Southeast AsiaOlsera2026-07-30

Olsera: Turning 80,000 Merchants' Transaction Data into Plain-Language Answers

Terabytes of transaction data across 80,000+ merchants sat almost entirely unused. Olsera built a merchant analytics chatbot on MOI, pairing a Medallion-architecture warehouse with NL2SQL for data questions and RAG for operational ones.

Olsera

Olsera is one of Indonesia's leading POS and business management platforms, serving more than 80,000 retail merchants across restaurants, retail stores and other formats, with terabytes of order, product, customer and payment data on the platform.

80,000+
Retail merchants served
Terabytes
Transaction data
NL2SQL + RAG
Dual answer paths
Row-level
Tenant isolation

The challenge

Olsera sits on a massive asset: Terabytes of transaction data across 80,000+ merchants-orders, products, customers, payments, channels-but none of this data was being surfaced back to merchants as actionable insight. Merchants had no self-service way to understand their own business performance: no sales trend analysis, no top-product rankings, no cross-store benchmarking. Olsera needed a platform that could monetize its data on the merchant side-turning raw transactions into a value-added analytics and AI service that merchants would pay for, accessible through a simple conversational interface requiring no technical skills.

The solution

Olsera deployed an AI Merchant Analytics Chatbot on OmniFabric, combining a Medallion Architecture data warehouse with an AI chatbot that merges NL2SQL (for transaction queries) and RAG (for POS knowledge base) to serve each merchant with personalized, access-controlled insights.

The outcome

Olsera transforms raw transaction data into a value-added AI service for merchants-turning a cost center into a new revenue stream with analytics and chatbot capabilities built on a single platform.

Each merchant asks business questions in natural language-sales totals, top products, order trends-and gets instant, data-backed answers through the AI chatbot, with no technical skills required.

A single chatbot handles both structured data queries (NL2SQL against transaction tables) and unstructured knowledge retrieval (RAG against POS help articles), eliminating the need for separate tools.

Solution Architecture

Data sources
  • Transaction data across 80,000+ merchants
  • Product, customer and payment data
  • Channel and store operations
  • Platform documentation and help content
MatrixOne Intelligence
  • Medallion-architecture warehouse
  • NL2SQL for data questions
  • RAG for operational how-to
  • Row-level security per tenant
  • Automatic merchant context
Business applications
  • Self-service merchant analytics
  • Sales and product insight
  • Channel comparison and trends
  • No analyst or SQL required

Related solutions and products