RoundCircle

About RoundCircle

We build Artha — India's specialised Finance LLM and Decision Intelligence platform. Artha is a controlled, finance-tuned large language model purpose-built for Indian BFSI — the foundational reasoning layer banks, lenders, and insurers build on. Unlike generic LLMs, Artha is fine-tuned to reason in the language of Indian finance: SEBI, RBI, IndAS, GST, and IRDAI. Generic foundation models are powerful at knowledge. But regulated financial decisions require domain expertise, regulatory fluency, institutional context, and auditability — layers no general-purpose LLM provides. 95% of enterprise GenAI pilots never reach production. In Indian BFSI, they stall because a global LLM can't be trusted in a decision a regulator will examine. Artha closes that gap — purpose-built, not prompted. We've built a specialised Finance LLM fluent in Indian regulatory and accounting language, three intelligence pillars: Business Finance, Tax, and Regulatory Intelligence, four knowledge layers: Foundation Role Institutional Outcome Intelligence, CIRCLE governance: explainability, data lineage, eval benchmarks, drift monitoring on every release, and a two-layer model: you build on our LLM, you own what you build. Your data never trains the shared model. Capabilities include credit underwriting, regulatory compliance, audit & reconciliation, portfolio intelligence, insurance product intelligence, tax assessment, and policy drafting. We serve banking, lending, capital markets, insurance, and government finance departments.