Box Unveils Skills and Skills API

San Francisco-based Box has just unveiled Skills and the related SDK, Skills Kit. With these new offerings, organizations and developers now have the ability to pull insights from their massive content stores in Box data sets and apply machine learning to release the intrinsic commercial value in that content.

Box Skills

The releases, unveiled at BoxWorks, are currently in preview and will soon enable those who store and manage their content in Box to apply third-party vendor algorithms to their content to resolve specific business problems.

To get the ball rolling, Box is previewing three initial Box Skills, using machine learning tools from Google Cloud and Microsoft Azure to solve common business use cases: Those use cases include:

  • Image Recognition: Detecting individual objects and concepts in image files, capturing text through optical character recognition (OCR), and automatically adding keyword labels to images to easily build metadata on image catalogs.
  • Audio Transcription & Analysis: Uses audio files to create and index a text transcript that can be easily searched and manipulated in a variety of use cases.
  • Video Indexing: Video Indexing analyzes video files to provide text transcription, topic detection and indexing, and facial recognition.

But according to Jeetu Patel, Chief Product Officer at Box, this is only the beginning. In the future, Skills framework will enable users to apply algorithms from any vendor to their content stored in Box. “What we aim to do is to take all those different machine learnings additions [that have been developed over the past 24 months] and bring it to the content that is in Box which will increase the value of your content in Box,” he said.

In addition to the initial Skills for image, audio and video analysis, Box will introduce Box Skills Kit, a set of developer resources that can be used in building custom Box Skills, the company said in statement.

Custom Skills can include chaining together multiple skills to enable intelligent business processes, leveraging a third-party machine learning solution (like form extraction) for a specific workflow, or training a skill to handle data unique to a business or industry use case.

The Content Challenge

The introduction of Box Skills responds to two basic ideas:

  • A piece of content in Box is infinitely more valuable than a piece of content outside of Box
  • The more content you put in Box the more value it will generate for you

Patel also points out that over the last year alone the amount of content in Box has doubled. Now, he said, there are 30 billion files stored in Box that instigate one trillion actions ever year.

While Box has been developing its cloud content management capabilities — Patel underlines cloud rather than on-premises — the only way to get the kind of insights Box customers are looking for is using machine learning and algorithms.

Patel’s insistence on the cloud comes from the fact that because data sets are getting so big, because the tools to manage that content are getting so dynamic, it is impossible to manage these problems anywhere other than the cloud.

Organization Problems

For organizations the problem is a simple one. They have huge amount of content, but they can only extract a small amount of the value contained in it. “The more content you have in the system the harder the system becomes to use — its inversely proportional in a sense. With most other industries the more content you have the easier your job is. In this industry, the reverse is true — the more content you have the more difficult it is to use," Patel said. “The way we think this can be solved is through machine learning technologies. Over the past 12 to 24 months there has been an increasing amount of developing in the machine learning space, not just from the start up vendors but even form the larger vendors. This is what we want to offer our users.”

Patel points to some of Box’ partners whose algorithms will be useable through Skills, notably Microsoft with Azure, IBM’s Watson as well as Google as an example of the kinds of intelligence enterprises might be interested in using. One useful use case examples Patel offers is video analysis.

Box Skills Video Analysis

Box Skills video analysis

In the above slide someone using an algorithm for video analysis can provide text, transcription, topic detection and indexing and facial recognition.

The only limit here is the number of algorithms that exist now and in the future, as the Skills Kit will allow developers to customize the Skills in ways that even the original algorithm provider hadn’t thought of. In fact, with Skills Kit users can even chain different algorithms together.

Users could, for example combine Microsoft Transcription and topic detection with IBM Watson’s Sentiment Analysis for a deep understanding of their audio content.

Machine Learning Paves The Way Forward

Box doesn’t want to limit themselves to their own vision of innovation but instead looks to tap into a deeper pool of the entire industry. “This is being driven by our core believe that we should not be limited by our own innovation velocity; we want to use the ecosystems development velocity,” Patel added. “The aggregate industry will always innovate faster than one company. We want to pull all that innovation together.”

When Will Box Skills Be Available

Box Skills and the Box Skills Kit will be available in public beta in early 2018.