Tel Aviv-based Factify has raised $73 million in a seed funding round to establish a new document standard for AI, the company announced Wednesday. The round was led by Valley Capital Partners, with participation from technology and business leaders, including John Giannandrea, former Head of AI at Google and SVP of AI at Apple, Ken Moelis, Founder of investment bank Moelis & Co., and Peter Brown, CEO of hedge fund Renaissance.
“We are building Document-as-Infrastructure,” says Matan Gavish, Factify’s founder and CEO. “Every document is a governed, intelligent asset that carries its own identity, access control, version history, and automation. Instead of relying on external systems to track who can access a file, which version is real, or what actions it should trigger, that intelligence and audit trail is integrated into the document itself.”
The Portable Document Format or PDF was introduced by Adobe Systems in 1993 and was released as an open standard in 2008. It has evolved over time, adapting to the internet and, later, to mobile phones. Will it survive the rise of AI, where data drives everything, including data and information locked in a PDF?
Adobe is responding with the Adobe Acrobat AI Assistant, which allows users to ask questions, summarize, and format documents directly within Acrobat. New competitors are using AI to seize on the opportunity to update how we work with documents.
“Universal, Productive, Delightful PDF” or UPDF, for example, offers AI tools for reading, editing text and images, annotating, organizing pages, scanning and securing PDFs; Smallpdf and Sejda allow users to use AI to edit text, rewrite content, and modify PDFs directly in a browser; and Wondershare PDFelement provides AI features for summarizing, translating, and editing documents. The interest in applying AI to documents is intense, and Andrew Ng’s DeepLearning.AI recently launched a new course, Document AI, teaching how “to build agentic document processing pipelines that extract structured data from PDFs and images while preserving layout, reading order, tables, charts, and forms.”
Gavish is not impressed. “Most of the new AI features around PDFs are assistants layered on top of a legacy PDF foundation,” he says. “They help users read, summarize, or chat with a document, but they do not operate on a definitive source of truth. The document itself remains static, disconnected, and fundamentally ungovernable. Companies still need multiple software solutions (and increasingly multiple AI assistants) to get things done.”
In contrast, “Factify’s vision is that all document work will be fully AI-based, not AI-assisted. To build this future, we are not adding AI to PDFs; we are updating what a digital document is.”
Gavish has an even broader vision of what he calls “the AI revolution.” He argues that our digital lives are now driven by a new form of large-scale cooperation among a community with the shared goal of developing better AI models, new models of participation, and new standards for evaluating output. “As AI systems begin to review, approve, share, and act on information, businesses need a new digital foundation their agents can trust,” says Gavish. “Today’s file-based digital document standard was never designed for this world. Documents today are static snapshots in a dynamic, automated environment.”
A “Factified Document” records every meaningful event in a permanent audit log and enforces governance directly within the document. As it is machine-readable and uniquely addressable, it remains continuously available as a verifiable source of truth for both people and AI systems. This allows organizations to maintain a single authoritative document that retains its identity, audit trail, and purpose wherever and whenever it is used.
Factify’s go-to-market strategy is initially aimed at regulated industries such as banking, insurance, legal services, human resources, and operations. “Our view is simple,” says Gavish. “AI will reshape how work gets done, but documents remain the backbone of regulated business and departments. For AI to be trusted with action, documents must evolve from files into infrastructure. That is the role Factify is built to play.”
