Home / Technology / Unsiloed Secures $7.5M for AI Document Clarity
Unsiloed Secures $7.5M for AI Document Clarity
18 Nov
Summary
- Unsiloed raises $7.5 million to make complex documents AI-accessible.
- Company uses dual-stream vision and heat map chunking for accuracy.
- Early customers include finance, legal, and healthcare sectors.

Unsiloed has successfully raised $7.5 million in seed funding, with Felicis spearheading the investment round. This capital injection is earmarked for enhancing their technology, which is designed to extract and structure information from complex business documents, making it readily usable by AI agents. The company addresses a critical gap where traditional AI struggles with the nuances of tables, charts, and varied document layouts.
Co-founders Aman Mishra and Adnan Abbas, formerly of IIT Kharagpur, developed a unique 'dual-stream' vision model and 'heat map-based chunking' technique. This allows for separate processing of text and image data, with focused computing power on crucial document areas. This innovative approach has demonstrated high accuracy, with tests showing a 99.78% success rate on financial documents.
The startup is already gaining traction with early customers across the finance, legal, and healthcare industries. Unsiloed's technology offers a competitive edge in a market with established players, promising more reliable AI interactions with crucial business information. The company plans to use the funding to expand its team and further refine its proprietary datasets.