Unsiloed AI secures $7.5M for document AI agents
Unsiloed AI raised $7.5 million in seed funding to develop technology that extracts structured data from complex documents for AI agents in finance, legal, and healthcare.

Unsiloed AI has raised $7.5 million in seed funding. The round was led by Felicis with participation from Y Combinator, Orange Collective, Uncorrelated Ventures, and angel investors Aarthi Ramamurthy and Beth Turner.
Co-founder and CEO Aman Mishra started the company with Adnan Abbas, a former classmate from IIT Kharagpur. They developed the technology before joining Y Combinator's Fall 2025 batch. The seven-person startup will use the capital to hire researchers and engineers, expand sales, and develop proprietary datasets.
Technology and performance
Unsiloed's software converts complex enterprise documents into structured data for AI agents. It specifically handles tables, charts, and detailed layouts that standard AI systems often misread. The technology also shows where each piece of extracted information appears in the original document for verification.
The startup uses a dual-stream vision model that processes text and image data separately. A heat map-based chunking system directs more computing power toward critical elements like figures and numbers. A specialized decoder accounts for industry-specific terminology to distinguish context-dependent terms.
According to the company, its system achieved 99.78% accuracy across 84,000 data points extracted from 542 financial documents. These results have not been independently verified. On the olmOCR-Bench, which tested 1,403 PDF pages with tables, math expressions, and complex layouts, Unsiloed's score outperformed key competitors.
| Company | olmOCR-Bench Score |
|---|---|
| Unsiloed | 88.0 |
| Reducto | 66.0 |
| Unstructured | 39.9 |
Customer applications
Unsiloed is already working with customers in finance, legal services, and healthcare. One bank is using the technology to process documents. A law firm has deployed the software on its own servers to handle contracts and case files.
Voice AI startup Bolna AI is also a customer. It uses Unsiloed to ingest knowledge bases for building voice agents.
Market positioning
The company operates in a crowded document AI market. Competitors include Reducto, Unstructured, Microsoft, and Amazon. Unsiloed argues that large language models are inherently probabilistic and prone to hallucination, which negatively impacts downstream applications. Its technology aims to reduce such errors in sectors like banking, healthcare, and law.
Pricing for the service starts at one cent per page. The funding will support the company's plans to hire researchers and engineers, expand its sales efforts, and develop proprietary datasets.





