Aaron Levie leads Box into its third era focused on workflow automation and AI | TechCrunch

Box began life almost 20 years ago in a dorm room at USC when Aaron Levie conceived of an online file storage and sharing system. A few years later, Levie’s original idea was becoming commoditized, and he switched gears to enterprise content management in the cloud. It was a radical notion at the time in an industry that was dominated by on-prem giants like Microsoft, EMC, IBM and OpenText.

Traditional enterprise content management, whether on prem or in the cloud, has involved storing, managing, securing and governing unstructured content. This has always been more difficult to handle than data sitting in neat columns and rows in a database.

Today, the industry is changing once more, and Box is again working to position itself on the forefront of that shift. Levie has always had a knack for seeing where the puck is going, and his company is embracing the software shift toward AI and workflow automation.

Last year, Box bought Crooze, a small company that specializes in workflow automation and metadata management with integrations into Box, making it a logical acquisition target. Being able to manage metadata is central to a lot of automation in content management because it provides a way for the software to identify and understand the type of content when there is no other structure present. That can help move different content types — whether documents, video, images or audio — through automated workflows and reduce a lot of monotonous tasks previously handled by bored and annoyed humans.

But what Box is doing with Crooze and generative AI may be part of a larger content management industry shift, one that could be as important as the move from on-prem to the cloud that Box helped lead 15 years ago.

Putting content to work

Levie is genuinely exuberant about the possibilities the Crooze technology can bring to the platform. “This is a very big deal. The way to think about it is that for the first time ever within Box, you’re going to be able to build no-code applications that let you render your content for any business process that you want,” Levie told TechCrunch. In other words, users can build custom applications that mirror business processes and make the content much more useful.

He recognizes that the folder structure can only get you so far, especially when dealing with large amounts of unstructured content like contracts, for example. It becomes unwieldy pretty quickly to try and find a contract, nevermind more-detailed pieces of the contract, when rifling through virtual folders.

“But with a no-code application development environment, you can build an actual dashboard that displays all of your contracts, all the data in those contracts and helps you automate the workflows around those contracts,” he said. That could involve editing, approvals, electronic signatures and so forth.

Generative AI plays a role here, too, letting users query the content in the folders to understand it better or locate specific pieces in a way that traditional enterprise search hasn’t been able to do. Summarization capabilities give users the gist of a large cache of content without having to read every line. In terms of workflow, generative AI’s coding capabilities can help build custom workflows based on particular requirements automatically.

It feels like Box is entering a new phase, says William Blair analyst Jason Ader, who watches Box. “Now I think we’re seeing Box 3.0, where it is moving into this AI and workflow realm and really going at the heart of a lot of those vertical industry workflows. These are tied to contracts and digital assets in obviously document-heavy types of industries where frankly AI has a massive role to play because it can automate a lot of that work,” Ader said.

Indeed, the way customers view content is changing. They don’t just want to manage it anymore, they want to put it to work in much the same way that data platforms like Snowflake and Databricks have moved beyond pure data management to building applications on top of it. Just having content sitting in storage repositories isn’t enough anymore, and AI is driving the push to automate workflows and produce practical business productivity outcomes.

“At the end of the day, enterprises want to leverage that content — not just store it — to drive automation and improve business outcomes,” said Alan Pelz-Sharpe, founder and principal analyst at Deep Analysis. “And hence acquisitions like Crooze provide ever simpler tools to develop those outcomes. Crooze is probably the most significant acquisition Box has made to date.”

Content management industry evolution

Box is hardly alone in this push, but as generative AI advances the ability to generate content and query the content store, we are starting to see content management and knowledge management (business memory) merging together. What’s more, the ability to generate code could allow companies to create custom workflows on the fly based on the requirements and types of content.

Cheryl McKinnon, a Forrester analyst who has been covering content for management for two decades, says she sees the content management industry as whole moving in the same direction as Box, and she believes it is a natural progression. “I see this is just moving up the maturity curve, and this shift towards workflow and AI is absolutely where the market has been moving,” McKinnon said. “This is kind of a turning point where now it’s not just about storing files and folders, but can we put that stuff to work? Can we think about content, not just from the storage point of view, but across the context of a whole business activity?”

This is a big moment for the whole industry, says Pelz-Sharpe. “The ECM sector as a whole (which includes Box) now has the biggest window of opportunity they have had in 20 years, opened for them by the interest and embrace of organizations large and small to leverage AI,” he said.

He thinks that ECM firms in particular are in a good position to take advantage of AI because they are already ensuring that unstructured data is accurate, relevant, secure and timely. That’s an important piece that AI models need that is often missing, he said. But the question is: Can Box and these other companies execute and take advantage of this moment?

“It’s important to note that although this window of opportunity is real, there is no guarantee ECM firms will pivot to embrace it,” Pelz-Sharpe said. “Firms like Salesforce, for example, are wising up to the importance of managing unstructured data, as is Oracle [and other industry giants].”

“Where Box and its ilk currently have an advantage, is that they have dedicated platforms to do this work, and equally importantly, deep skill sets and experience to bring to the table.”