Skip to main content
6h ago

Unsiloed AI Secures $7.5M to Bridge the Gap Between Complex Documents and Intelligent Agents

As businesses increasingly rely on AI agents for automation, the challenge of accessing data hidden within complex documents remains a significant hur

3 min read9 views5 tags
Originally reported bytechcrunch

As businesses increasingly rely on AI agents for automation, the challenge of accessing data hidden within complex documents remains a significant hurdle. Unsiloed AI has secured $7.5 million in seed funding led by Felicis to address this issue, aiming to enable enterprises to integrate their document repositories with AI agents more effectively. Silicon Valley venture capital firm Felicis spearheaded the round, joined by Y Combinator, Orange Collective, Uncorrelated Ventures, and angels Aarthi Ramamurthy and Beth Turner.

Financial and legal documents frequently contain tables, charts, and intricate layouts that can confuse standard AI systems. Inaccuracies in interpreting this data can result in significant financial losses, a risk that is particularly acute for sectors like banking, healthcare, and law, where precision is paramount.

Unsiloed provides a software solution designed to deconstruct these complex files. By extracting data from tables, charts, and other elements, the platform converts unstructured information into structured data that AI agents can utilize. Additionally, the technology traces the source of the extracted data back to the original document, allowing businesses to verify the results rather than accepting AI outputs blindly.

"The fundamental issue with large language models is their probabilistic nature," explained Unsiloed co-founder and CEO Aman Mishra. "There will inevitably be deviations and hallucinations, which can have severe downstream consequences."

Mishra founded the company with Adnan Abbas, his former classmate from the Indian Institute of Technology (IIT) Kharagpur. Both joined Y Combinator's Fall 2025 batch after developing the technology.

Unsiloed already serves clients in finance, law, and healthcare. Mishra revealed that a bank uses the technology for document processing, and a law firm has deployed it on-premises to manage contracts and archives. He also mentioned that voice AI startup Bolna AI is a client, using the tool to feed knowledge bases into voice agents.

The San Francisco-based startup faces a crowded market, competing with specialized firms like Reducto and Unstructured, as well as offerings from Microsoft and Amazon. General-purpose AI models also pose a challenge in this space.

According to Mishra, Unsiloed distinguishes itself through its handling of complex documents. The startup utilizes "dual-stream" vision models to process text and image data separately, alongside a "heat map-based chunking" method to highlight key document sections.

"One stream handles text tokens while the other manages exact image tokens," Mishra described. "We have invested considerable effort into architecting this specific approach." The system then allocates more computational resources to figures and numbers. Furthermore, a specialized decoder accounts for industry-specific jargon to prevent misinterpretation of terms with multiple meanings.

To validate its claims, Unsiloed cited a test involving 542 financial documents, where it achieved 99.78% accuracy across 84,000 data points. This compared its performance against Microsoft's Azure Document Intelligence and the client's internal tools.

The startup also reported outperforming rivals on the olmOCR-Bench public benchmark. In its tests, Unsiloed scored 88.0, surpassing Reducto's 66.0 and Unstructured's 39.9. The benchmark assessed the systems' ability to handle tables, mathematical formulas, scanned images, and complex layouts across 1,403 PDF pages.

Pricing begins at one cent per page. Mishra noted that the targeted approach helps reduce costs by focusing computing power on critical sections of a document rather than uniform processing. Currently, Unsiloed has a team of seven. The company plans to use the new funding to expand its research and engineering teams, scale its sales operations, and develop proprietary datasets to enhance its technology.

#AI News#Document processing#Enterprise AI#Dual-stream vision#Data extraction
ES
Editorial StaffEditor

The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.

View all posts
Reader feedback

What did you think of this story?

User Comments

Filter:
No comments yet. Be the first to comment!
Continue reading
View all news