Skip to content
worth noting Business and investment

Startup Unsiloed AI raises $7.5 million for document processing for AI agents

only one source so far

Startup Unsiloed AI raised $7.5 million in a seed round led by Felicis for software that turns complex corporate documents into data usable by AI agents. According to the company, the technology achieves high accuracy, but the results have not been independently verified.

Startup Unsiloed AI announced seed funding of $7.5 million, led by venture capital firm Felicis. The round also included Y Combinator, Orange Collective, Uncorrelated Ventures, and angel investors Aarthi Ramamurthy and Beth Turner. The company was founded by Aman Mishra (CEO) and Adnan Abbas, classmates from the Indian Institute of Technology Kharagpur, who went through Y Combinator's Fall 2025 batch.

Unsiloed AI develops software that breaks down complex documents — tables, charts, messy layouts — and converts them into structured data usable by AI agents, while showing which part of the original document each extracted piece of data comes from. According to the company, this solves the problem of large language model hallucinations when processing documents, which can have serious consequences, for example in banking, healthcare, or legal services. According to the company, the technology uses "dual-stream" visual models that separate text and image tokens, and a "heat map-based chunking" method to identify the most important parts of a document, supplemented by a decoder adapted to industry-specific terminology.

The company states that in a test on 542 financial documents from a customer, it achieved an accuracy of 99.78% out of approximately 84 000 extracted data points, outperforming the compared large AI models, Microsoft's Azure Document Intelligence tool, and the customer's internal solution. On the public olmOCR-Bench benchmark from Allen AI (1 403 PDF pages), the company reports a score of 88.0, compared to 66.0 for Reducto's agentic system and 39.9 for Unstructured. According to the article, these results have not yet been independently verified.

According to the company, Unsiloed AI's customers include a bank, a law firm running the tool on its own servers to process contracts and old cases, and voice AI startup Bolna AI, which uses the tool to process knowledge bases. Pricing starts at one cent per page. The company has seven employees and plans to use the new funding to hire researchers and engineers, expand sales, and develop its own datasets. It competes with specialized startups Reducto and Unstructured as well as offerings from Microsoft and Amazon.

What changed

Why it matters

Companies in industries that depend on accurate document processing (banks, law firms, healthcare) are gaining another specialized player in the data extraction market, positioning itself against general-purpose AI models as well as large cloud offerings. The stated accuracy and benchmark figures come solely from the company itself and have not been independently confirmed, which should be taken into account when deciding on deployment.

Relevant practical impact

What this means

01

For a business

A funded competitor has entered the market for tools that extract data from corporate documents (tables, charts, complex layouts), with pricing starting at one cent per page and deployments among clients in banking, legal services, and healthcare — companies in these sectors now have another documented option to compare, although the benchmark results have not yet been independently confirmed.

Productivity
What to decide When choosing a tool for extracting data from corporate documents, include Unsiloed AI among the tested options alongside Reducto, Unstructured, Azure Document Intelligence, and tools from Amazon, bearing in mind that the figures on…
More business impacts →
AI agents Document Processing Felicis seed funding Unsiloed AI venture capital

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
TechCrunch AI independent context · first detected Felicis leads $7.5M bet on Unsiloed AI to make enterprise documents usable by AI agents