Box has today announced it will be introducing machine learning to its cloud storage and content management platform. Speaking at the company's annual conference BoxWorks in San Francisco, CEO Aaron Levie announced that the new capabilities will use artificial intelligence to help customers extract more value from their data.
The company boasts 76,000 customers, including 65% of Fortune 500 companies, and those organisations are storing vast quantities of information on the platform. According to Box's chief product officer Jeetu Patel, the total amount of customer data hosted by the company is doubling every year, with companies opting to store their most sensitive and mission-critical information on the platform, in addition to storing richer media like images, video and audio files.
However, Box doesn't just want to provide storage and hosting for its customers. Rather than just being somewhere to dump their data until they need it, Box wants to add value for their customers. The biggest problem with this is that as the size of a company's data archives balloons, it becomes harder and harder to manage and organise it, let alone extract actionable insights from it. "The more content you have, the harder it is to make sense of it," Patel told IT Pro; "that's just the nature of the beast."
For Patel and his team, machine learning tools offer the ideal way to deal with the vast amounts of information held by their customers. "We think that machine learning is probably one of the most scalable ways to go out and solve this problem," he said. Unfortunately, while there are a battery of clever machine learning systems that can perform tasks like automatic audio transcription and object recognition within images, there's currently no way for companies to apply them to their content at scale.
Enter the newly-unveiled Box Skills framework, a method for applying these machine learning capabilities to their Box-hosted content. Box is hugely excited about this announcement - Patel referred to it as "arguably one of the biggest announcements we've made in the history of the company".
"We are in the midst of a revolution in enterprise software driven by artificial intelligence and machine learning, and we are making Box the most intelligent cloud content management platform in the world," Levie said. "As businesses continue to drive digital transformation, they need to realize more value and intelligence from their content. Box Skills is a first-of-its-kind framework that will make it possible to digitize almost any business process on Box."
The new Box Skills will be launched in public beta early next year, with pricing still to be announced.
As part of the reveal, Box demonstrated three Skills that are currently in development; an audio skill, a video skill and an image skill. The audio tool, based on IBM's Watson technology, offers automatic transcription of audio files. The video skill uses Microsoft Azure's Cognitive Services to analyse video clips and provide transcription, along with topic detection to tag the contents of video and facial recognition tracking which frames the participants appear in. Finally, the image recognition tools analyses images through OCR text recognition and object detection to automatically tag pictures with relevant metadata, using Google Cloud Platform.
"IBM has shown how data-intensive industries are being transformed through the use of Watson," said David Kenny, senior vice president for IBM Watson and Cloud Platform. "Now we're enabling enterprises across the globe to combine the rich content in Box with the AI power of Watson to reimagine critical business processes and how they get work done."
"It's fantastic to see that Box is starting to introduce artificial intelligence directly into their platform," said Constellation REsearch analyst Alan Lepofsky. "The way they are doing it is by leveraging other AI Frameworks from companies like IBM, Google and Microsoft. This allows them to be vendor-agnostic, and use the best features from various companies."
The potential use-cases are immediately apparent; a creative agency, for example, can instantly browse through all of its art assets to find any related to a given activity or profession, while a speech made by the CEO of a multinational company can be automatically transcribed and translated for employees in other regions with no extra time or effort.
"If you think about the fact that organisations today spend millions of dollars in doing manual data entry to classify and tag documents and make sure that they can populate metadata against those documents, there is definitely an efficiency play over there by making sure that you can use these capabilities, so you don't have to do that at scale," Patel told IT Pro.
Forrester analyst Rob Koplowitz agreed that integrating AI into Box's platform was a good move. "The enterprise use cases for image, video and audio recognition are beginning to emerge," he said. "Adding that to the place where my content is stored can be very compelling. They were smart to leverage external AI service providers. It will require some due diligence on the part of customers to determine if any privacy, security, compliance or auditability issues are introduced but for most organizations this will represent a very smart approach on the part of Box."
