What’s inside Box? More generative AI

Box got its start a decade ago as a cloud based document sharing technology. The platform has evolved significantly over the years to become an enterprise content cloud, and today is taking another step forward with the launch of Box AI.

With its new service, Box is providing a set of generative AI capabilities that will enable organizations to better understand and create new enterprise content. At launch, Box AI will benefit from an integration with OpenAI’s large language model (LLM) services, with the plan to have additional LLM providers over time.

This new move is not the first time Box has had AI on its platform. Back in 2017, the company announced a partnership with Google to integrate AI for image recognition. In 2021, Box tapped deep learning to help with security efforts and detect sophisticated malware.

“What we’re announcing with Box AI is a broad platform where you’re going to be able to use AI to generally work with content and understand it in new ways,” Aaron Levie, cofounder and CEO of Box, told VentureBeat. “It’s really driven by these LLMs that have so much more breath to them in terms of what kind of problems we can solve.”

How ChatGPT inspired Box (and many others)

Box AI is just the latest in a string of enterprise software vendors that have added some form of generative AI to their platforms.

Microsoft has been steadily adding what it refers to as ‘copilot’ capabilities to its enterprise software services to enable users with AI powered functionality. Salesforce has similarly expanded its enterprise software-as-a-service (SaaS) platforms with generative AI under the banner Einstein GPT.

For Microsoft, Salesforce and now Box, a key driver for integrating generative AI is the runaway popularity of OpenAI’s ChatGPT.

Levie said that his team was playing with ChatGPT when it first came out at the end of 2022, and within a day realized the huge impact and potential it had for IT generally — and for Box’s users, too. Box’s executive management saw that they could use the power of generative AI to gain insights from any type of document, which could be a real benefit to the company’s enterprise users, Levie explained.

“We realized there were so many use cases where we could just make work more efficient and, in many cases more delightful, and really get the entire power of LLMs brought to enterprise content,” said Levie.

What’s inside Box AI

Box AI is not training a new AI model — rather it is bringing the power of existing LLMs to data that an enterprise has on the Box platform.

Instead of a generic search to look for specific keywords inside of a document, Box AI will now allow users to ask questions about what’s in a document. Levie said this allows enterprise employees to interact with their content as a source of knowledge that AI can learn from, provide analysis and helping organizations to be more productive.

For example, Levie said that a user could be looking at a budget document and ask the AI to help brainstorm ways to improve  processes or save money. Another example could be a user looking at a legal contract and asking what the riskiest clauses are in the contract? 

“Being able to work with your data and ask new kinds of questions on top of your content — that’s what Box AI is,” said Levie.

Generating new content

The platform also includes the ability to generate new content with Box Notes, which is Box’s online document editor and collaboration workspace. Levie said that now with Box Notes, an enterprise will be able to use Box AI to write a meeting agenda, draft a blog post and summarize information or any other type of content.

“You can imagine that Box AI will be plugged into a variety of the components of our platform,” said Levie. “It’s really about turning our platform into an intelligent content cloud that helps you unlock the value of your content.”

Overall, Levie expects to see the power of generative AI and intelligence infused across all enterprise software vendor offerings in the future as simple table stakes.

“When you have a platform shift like this, it becomes a fundamental requirement from your customers and you have to participate so it really turns into table stakes,” said Levie. “I don’t think there’s an enterprise on the planet that would accept that their enterprise software in five years from now is just not intelligent.”

Originally appeared on: TheSpuzz