Box acquires Alphamoon intelligent document processing tech

Box Inc. has acquired Alphamoon to further its intelligent document processing technology and enterprise knowledge management AI platform.

IDP adds AI on top of traditional optical character recognition (OCR) to scanned paper document images or unstructured PDFs.

AI such as natural language processing (NLP), workflow routing and document structure recognition has been around for years. But Alphamoon adds generative AI (GenAI) on top of those existing versions of AI.

Rand Wacker, vice president of AI product strategy at Box, explained that adding GenAI aids in not only summarizing documents but also extracting content from documents. GenAI can also recognize the structure of documents, categorize them, and work with existing OCR and NLP to make digital conversions of paper more accurate.

Technically, Box hasn't bought a large language model but a toolkit that will enhance the Box AI platform. But Box AI incorporates retrieval augmented generation that commingles a Box user's own content with outside LLMs to keep data secure and, moreover, train Box AI to better understand what type of document an incoming file is. Alphamoon adds sophistication to that process and, in concert with other Box features, will enable admins to create tools to accomplish it more easily.

"Say, for example, if Alphamoon OCRs something and it gets something wrong or the extraction's wrong. It can tweak that one part, and it goes back into a [large language] model," Wacker said. "So what it does is actually powered by an LLM. But the things that it's being trained on [teach it] how to understand these documents rather than historical content on the internet."

Previewing an upcoming Deep Analysis report, founder Alan Pelz-Sharpe said that respondents to a survey of 500 enterprises in vertical industries such as financial services, manufacturing, healthcare and government reported that 53% of enterprise documents exist on paper. That means Box users need more accurate tools to get those contracts, letters, invoices, faxes and remaining piles of paper digitized more accurately.

"IDP is incredibly difficult to do, but AI has made huge advances in doing it," Pelz-Sharpe said. "If a piece of paper is crumpled, if there's a coffee stain on it, if it's handwriting -- I'm not saying everything works perfectly -- but if something is slightly illegible, AI's moved the accuracy way forward. … This was a gap for Box. They had relied on partners; it was always a bit of an Achilles' heel."

Wacker said Box plans to offer Alphamoon tools in its platform later this year, with deeper integrations next year for no-code app building related to Crooze, another acquisition, as well as with Box Relay's forms and document generation tools.

Financial terms of the acquisition were not disclosed.

Don Fluckinger is a senior news writer for TechTarget Editorial. He covers customer experience, digital experience management and end-user computing. Got a tip? Email him.