Box announced on Thursday that it had acquired Alphamoon , a small Polish startup that specializes in intelligent document processing, a technology that helps companies extract metadata from dense text documents — think contracts and loan applications — automatically.
The companies didn’t share the purchase price, but Box CEO Aaron Levie said the acquisition provides a crucial piece of missing functionality for Box AI, the company’s underlying AI technology, to understand documents better and move them into workflows.
“Alphamoon has been building the technology to solve this problem of how do you connect leading AI to the underlying document, and then all of the management of the document extraction process,” Levie told TechCrunch. This involves extracting the metadata from a complex document, something it would take humans quite a while to compile, and putting it to work across the Box content repository.
This ability to extract metadata from text documents is called intelligent document process (IDP). “IDP is all the technology that involves taking in a document — a contract, an invoice, a financial record — and extracting the appropriate metadata from that document. And then obviously storing that in the metadata system and then letting you automate workflows around it,” he said.
Levie says that the acquisition works hand-in-glove with the Crooze acquisition earlier this year. “The two parts you have to have for modern intelligent content management is you have to be able to extract the metadata from the content in an intelligent way, and then you have to be able to actually render and automate workflows around it in a very simple, no code way as well.” The former is what Alphamoon does, and the latter is what Crooze handles.
After the company bought Crooze, it had been relying on third-party partnerships to handle IDP. When it looked at the build-versus-buy equation, it became clear that this was a specialized technology, and Box began exploring acquisition targets before settling on Alphamoon.
“How do you appropriately read a complex document from a structure standpoint? How do you then have a set of interfaces to help you manage that whole process and then connect to the Box AI platform to actually then leverage the underlying AI models? So that’s when we got really excited by Alphamoon technology,” Levie said.
This is not an easy problem to solve, says Alan Pelz-Sharpe, founder and principal analyst at Deep Analysis. “Unstructured data — PDFs, for example — are designed for humans to read them, not machines. So, in that regard, IDP is not just about capturing an invoice and processing it. It’s about converting all the unstructured files and making them ‘AI ready,’” he said. “And Box has a huge amount of data ready to be leveraged on its platform. So for me, this, in the short term, fills a gap, but in the long term gives them the opportunity to become an AI-first content platform.”
Alphamoon was founded in 2017 in Poland. It raised a modest €2.3 million, per Crunchbase .
While the acquisition has been finalized, the company did not disclose the purchase price or the number of employees coming over in the deal.