Unsiloed AI, backed by a $7.5 million seed investment led by Felicis, is pioneering a dual-stream AI approach to extract and structure data from intricate business documents, enabling safer and more effective use by enterprise AI agents.
- Raised $7.5M seed funding led by Felicis with Y Combinator and angels
- Uses dual-stream models to handle text and image data separately
- Claims 99.78% accuracy on complex financial document data extraction
What happened
Unsiloed AI announced a $7.5 million seed funding round led by venture capital firm Felicis, with participation from Y Combinator, Orange Collective, Uncorrelated Ventures, and angel investors. The company aims to improve how AI agents access and use data locked in complex enterprise documents by developing technology capable of handling diverse document elements such as tables, charts, and mixed layouts.
Co-founded by Aman Mishra and Adnan Abbas, both alumni of the Indian Institute of Technology Kharagpur, the startup has already engaged customers in finance, legal, and healthcare sectors. These clients use Unsiloed’s solutions to process sensitive documents with enhanced accuracy and usability, including a bank, a law firm, and a voice AI startup leveraging the tool to build knowledge bases.
Why it matters
Traditional large language models often struggle with the non-text components of business documents, which leads to hallucinations or errors in data extraction that can have costly downstream consequences, especially in regulated sectors such as banking, law, and healthcare. Unsiloed AI’s approach addresses these challenges by employing 'dual-stream' vision models that separately process text and image data in documents, increasing reliability and interpretability.
The ability to map extracted data back to its origin in source documents adds a layer of transparency critical for businesses that must verify AI outputs. A technical brief from Unsiloed cited a 99.78% accuracy rate across 54,000 data points from 542 financial documents, highlighting its potential advantage over existing tools from tech giants and startups in a competitive AI document-processing market.
What to watch next
Key developments to monitor include Unsiloed AI’s ability to expand its client base across other high-stakes industries and demonstrate independent verification of its accuracy claims. The startup’s competitive positioning against established cloud providers and specialized players will also be critical to its growth trajectory.
Additionally, how Unsiloed refines its dual-stream and heat map-based chunking technologies could set new standards for document processing by AI agents. Observers should also watch for potential partnerships or integrations with broader enterprise AI platforms, which could accelerate adoption and influence market dynamics.