document-ai
(24)
Rossum Taught the Machines to Read the Paperwork Nobody Could Kill
The Czech-born document-AI company made peace with the invoice, trained a specialist model to understand it, and sold the result to enterprises drowning in exceptions. Coupa’s 2026 acquisition turns that stubborn paperwork layer into part of a much larger spend-management machine.
Jeong Kim Built a PDF Company in a Sushi Shop. Then He Bought His Old Employer.
He left Artifex to turn printing technology into paperless software. Eighteen years later, he returned as the buyer - and now wants PyMuPDF to become a gateway between the world's documents and its AI models.
Alberto Gimeno Learned to Sell the Software He Used to Build
Three university friends spent a decade building together. Their useful discovery was not a flashier model, but a stubborn lesson: find the market that pulls, then give it your full attention.
Etienne Bernard Is Teaching AI to Read the Fine Print
The physicist who helped make machine learning automatic at Wolfram is now shrinking document AI into focused models that turn unruly files into useful data - with privacy, precision, and a well-placed null.
Aman Mishra Is Teaching AI to Read the Fine Print
After low-latency trading systems and enterprise copilots, the Unsiloed AI co-founder found his company in a broken table: the place where flat text loses the plot.
Datatruck Wants to Be the Operating System Trucking Runs On
A Chicago-founded startup is betting that trucking's back office - the spreadsheets, the phone tag, the paperwork - is a software problem AI can finally solve. About 500 carriers are already running on it.
Adit Abraham Found a Business in the Documents Everyone Else Hated
A failed memory tool, a two-week pivot and 154 days of enterprise courtship taught Adit Abraham a durable lesson: the useful work is often the work nobody wants to brag about.
6Estates
6Estates is a Singapore-based enterprise AI company that turns messy, unstructured financial documents into structured decisions. Spun out of the NExT++ research center (NUS, Tsinghua, University of Southampton), it builds domain-specific large language models, intelligent document processing, and agentic AI tools that banks and lenders across Asia use to automate credit assessment, financial statement analysis, and trade finance checks.
NuMind
NuMind is a Paris-based, Y Combinator-backed AI company building multimodal large language models that turn messy documents - PDFs, scans, spreadsheets, handwritten forms - into clean, structured JSON. Its NuExtract platform and open-source models let developers and enterprises extract exactly the fields they define, in any language, either as a hosted API or deployed privately on their own infrastructure.
The Boring Industry Quietly Becoming a Trillion-Dollar Category
Intelligent document processing is the unglamorous layer under every regulated industry - the plumbing that turns the mortgage packets, claims files and bank statements no computer could read into data a compliance team will stake a lawsuit on. Reading the page was never the hard part; proving it was read correctly is. As sovereign wealth funds and top-tier VCs write nine-figure checks into 'document processing,' three tribes are fighting over the same ground, and the real prize is not the document. It is permission for software to make the decision a human used to sign.
Unsiloed AI
Unsiloed AI builds an API that turns messy, multimodal documents - PDFs, scans, spreadsheets, charts and images - into structured JSON and LLM-ready Markdown. Founded by IIT Kharagpur alumni Aman Mishra and Adnan Abbas and backed by Y Combinator (Fall 2025), the company uses proprietary dual-stream vision models to parse, extract and split documents at scale, and ranks #1 on the public olmOCR-Bench, ahead of models from OpenAI, Anthropic and Google.
Hyperbots Inc.
Hyperbots is an agentic AI company building finance and accounting co-pilots for mid-market enterprises. Its platform automates core workflows - procure-to-pay, order-to-cash, expense management and analytics - by converting unstructured financial documents into structured, ERP-ready data. Founded in 2023 and headquartered in Dover, Delaware with engineering in Bengaluru, the company reports 99.8% document accuracy and up to 80% straight-through processing, and in 2025 raised a $6.5M Series A co-led by Arkam Ventures and Athera Venture Partners to expand in the U.S. and launch HyperLM, a large language model trained specifically on finance data.
Kushal Byatnal
Kushal Byatnal is the co-founder and CEO of Extend, a New York company building a full-stack cloud for document processing. Extend turns messy PDFs, images, and emails into production-ready structured data using vision models, large language models, and human-in-the-loop tooling, and is used by teams at Brex, Chime, Flatiron Health, Opendoor, and Checkr. A two-time technical founder and Duke computer science graduate, Byatnal was an early engineer at Brex and co-founded the creator fintech Stir before starting Extend in 2023. In June 2025 the company raised $17 million led by Innovation Endeavors.
Anvil
Anvil is a San Francisco software company that turns paper-heavy back-office processes into programmable digital workflows. Its API-first platform bundles PDF generation and filling, no-code webforms, embedded e-signatures (Etch), and workflow automation so developers and product teams can collect data, generate compliant documents, and get them signed without stitching together multiple tools. Founded in 2018 by Mang-Git Ng and Ben Ogle, Anvil serves customers across insurance, financial services, HR, real estate, healthcare, and education, and has raised $10M in Series A funding.
Sensible
Sensible is a developer-first platform that turns any document - PDFs, scans, emails, images - into structured, verifiable JSON. It pairs LLM-based extraction with deterministic, layout-based rules written in its own configuration language, SenseML, and ships 150+ prebuilt parsers plus confidence scores, source coordinates, and human-review workflows. Founded in 2020 and backed by Craft Ventures, Sensible is used by companies across insurance, proptech, financial services, logistics, and healthcare to automate document-heavy operations.
Mindee
Mindee is a Paris- and San Francisco-based AI company building an API-first platform that turns any document - invoices, receipts, passports, bank statements - into clean structured JSON. Founded in 2018 and backed by GGV Capital, Alven and Y Combinator, it gives developers pre-trained OCR endpoints plus a no-training-required custom builder called docTI.
Cyrus Khajvandi
Cyrus Khajvandi is the co-founder and CEO of Humata, an Austin-based AI startup often described as 'ChatGPT for all your files.' A Stanford biology graduate and serial entrepreneur, he built Humata after struggling to keep up with scientific papers, turning that personal frustration into a tool that lets people ask questions of their documents and get cited, fact-checkable answers. The company launched in early 2023, grew to millions of users processing tens of millions of pages, and raised $3.5M led by Google's Gradient Ventures with ARK Invest and M13.
Burmester & Vogel
Burmester & Vogel is a Boston-based maritime technology company that uses AI and machine learning to automate laytime calculation and demurrage claim management for the shipping industry. Founded in Hamburg in 1983 as a traditional laytime firm, it has reinvented itself as the maker of the Demurrage AI Copilot - software that digitizes shipping documents, calculates delay time, and audits demurrage claims in seconds. It serves industrial producers, commodity traders, bulk terminals and shipowners across several hundred companies worldwide.
H2O.ai
H2O.ai is a Mountain View-based AI cloud company founded in 2012, on a mission to democratize artificial intelligence for everyone. The company offers a converged platform spanning predictive and generative AI - from its open-source H2O-3 machine learning framework to enterprise products like Driverless AI, h2oGPTe, and the H2O AI Cloud. Trusted by over 20,000 organizations including more than half the Fortune 500, H2O.ai is known for employing the world's largest team of Kaggle Grandmasters and for pioneering automated machine learning. With $246 million in total funding at a $1.7 billion valuation, the company serves industries from financial services and healthcare to government and telecommunications, delivering AI that organizations can deploy on their own terms - on-premises, in the cloud, or in air-gapped environments.
Upstage
Upstage is a South Korea-born, US-expanding AI company building Solar - a family of frontier large language models - alongside Document AI tools that turn messy PDFs, scans, and contracts into structured data for enterprise workflows.
Shawn Heide
Shawn Heide is a Boston-based entrepreneur who blends culinary creativity with digital storytelling and enterprise AI. As co-founder of 9muse Creative Studio, he crafts cinematic brand films for companies across New England. His culinary arc spans founding Wagwan Jerk Bar in 2017 — a Caribbean-meets-New England pop-up launched at Boston farmers markets — to directing culinary operations at The Kitchen at White Barn, a Neapolitan pizza restaurant grown from a working farm. Pursuing his MBA at Babson F.W. Olin Graduate School of Business, Shawn is also connected to Upstage AI, the enterprise AI company known for document intelligence and large language models that raised a $120M Series C in early 2026.
Sung Kim
Sung Kim (김성훈), known professionally as Hun Kim, is the Co-Founder and CEO of Upstage AI, South Korea's first generative AI unicorn. A former HKUST professor and founding head of Naver Clova AI, he co-founded Upstage in October 2020 to solve the gap between enterprise data and practical AI deployment. Under his leadership, Upstage developed Solar - a compact 10.7B parameter language model - and Solar Pro 2, which outperforms GPT-4.1 at half the compute cost. With $326M in total funding and a Series C close in early 2026, Upstage is positioning itself as a global AI frontier company focused on enterprise document intelligence and AI agents.
Lucy Park
Lucy Park (Eunjeong Park) is the co-founder and US CEO of Upstage, the first Korean generative AI startup to achieve unicorn status. A PhD in Data Mining from Seoul National University and MIT Innovators Under 35 honoree, she built KoNLPy - the go-to Korean NLP library used in classrooms and production systems across Korea - before leading the Papago machine translation team at NAVER. At Upstage, she drives the company's AI models and document intelligence products into enterprises across insurance, finance, legal and healthcare, with the company's Solar LLM powering mission-critical workflows globally.
Dan Ashton
Dan Ashton is the CEO and Co-Founder of Brain Co., a San Francisco-based AI company that builds applied intelligence platforms for the world's most complex institutions - from government permit offices to hospital systems. He brought his AI startup Serene AI into Brain Co.'s orbit and now leads a company backed by $30M in Series A funding from Elad Gil, Affinity Partners, Patrick Collison, Reid Hoffman, and Andrej Karpathy. Before Brain Co., he was an early operator at three unicorns - Opendoor, Faire, and Clubhouse - and a McKinsey consultant before that. Brain Co. has already deployed AI that unlocks $375M annually in permitting value and mitigated $100M+ in supply chain losses.