Box Announces Expansion of Box Skills Private Beta to Bring Advanced Machine Learning to Enterprise Content

By STS News Desk On Jun 29, 2018

Introduces custom Box Skills from IBM and support for new machine learning services from Microsoft Azure

Box, a leader in cloud content management, today announced the expansion of its private beta program for Box Skills – the breakthrough framework for applying state-of-the-art machine learning technologies, such as computer vision and video indexing, to content stored in Box. Previously only available to select customers, Box Skills will be expanding its private beta access to additional customers starting in July, with several customers added each week.

Box also announced the availability of a new service offering from IBM to build custom Box Skills as well as Box Skills Kit support for new cognitive services from Microsoft Azure.

“Content is the next great untapped resource of insights and innovation in the enterprise”

Jeetu Patel

“Content is the next great untapped resource of insights and innovation in the enterprise,” said Jeetu Patel, Chief Product Officer, Box.

Jeetu added, “Box Skills brings ground-breaking machine learning to enterprise content for the first time and will enable organizations to reimagine and digitize business processes on Box. We’re about to enter the next major frontier of computing with AI and it will completely change how we work together and collaborate on content.”

At BoxWorks 2017, Box previewed three initial skills that leverage machine learning services – like IBM Watson, Microsoft Azure, and Google Cloud – to organize, protect and automate processes around business content in Box. The three initial skills include:

Image Intelligence: Detects individual objects and concepts and recognizes text in image files.
Audio Intelligence: Transcribes and identifies key topics in spoken audio files.
Video Intelligence: Transcribes and identifies key topics for speech and detects individual faces as they appear in video files.

Over the past few months, Box has on-boarded an initial set of customers, including Virgin Trains, Ancestry.com, The University of Chicago, City of San Jose and others, into the private beta program for Box Skills. The customers participating in the private beta are leveraging Box Skills for a wide variety of enterprise use cases and digital business processes. For example, a large insurance company is building a custom skill to automatically label household objects in images and videos as part of a digital homeowner’s insurance policy process, and a construction equipment rental company is building a custom skill to automatically identify equipment type and extract part numbers from vehicles in images and videos.

John Sullivan

“At Virgin Trains, we’re constantly looking for new ways to become a more digital business by accelerating processes and improving our employee and customer experiences,” said John Sullivan, CIO and Project Director at Virgin Trains.

John added, “With Box Skills, we can easily bring the best of artificial intelligence and machine learning to the way our organization works every day, helping us streamline collaboration and digitize business processes faster than we were able to before. We’re incredibly excited about the possibilities of Box Skills.”

Enriching Box Content with Custom Solutions from IBM Watson

Last week, IBM and Box announced the availability of a new service offering from IBM to build custom Box Skills that apply Watson AI technologies to the Box Skills framework. These new, customizable solutions, built using the Box Skills Kit, will enable businesses to work hands-on with the IBM team to apply IBM’s enterprise-strength Watson AI technologies to enterprise content managed in Box.

IBM introduced two new services for building custom Box Skills —

Custom image insights with Watson Visual Recognition

This custom solution analyzes image data, enriching it with classifiers to make it easy to search and consume, and training custom visual models that best address business needs. For example, an environmental organization could leverage a custom skill that can analyze satellite images of coastal erosion in Box, quickly detecting areas of most impact, speeding up the time to taking action and reducing the costs of monitoring.

Custom document insights with Watson Natural Language Understanding

This custom skill automatically tags documents stored in Box with relevant concepts, entities, and keywords to enrich the content and make it more searchable. This skill will also help users quickly access specific information in dense documents like research papers, service manuals, and legal documents. For lawyers working on time-sensitive matters, this could be the difference between making a court filing deadline and needing to file for a costly extension.

Support for the Latest Azure Cognitive Services from Microsoft

Designed as an open and flexible framework, the capabilities of Box Skills will grow and evolve as new and expanded machine learning services are made available from providers around the industry. To that end, Box also announced today support for new AI services available from Microsoft Azure that were announced as part of Microsoft Build 2018, Microsoft’s annual developer conference, including:

Improved OCR in Computer Vision

Using the improved OCR engine now available through Computer Vision, part of Azure Cognitive Services, any image-based Box Skill will now be able to better identify text within images, improving accuracy and speed of image-based workflows, such as digital asset management.

Expanded object detection in Computer Vision

Computer Vision is now able to detect dozens of more objects within images, making it easier to apply rich metadata to images in Box. Now, when images are uploaded to Box and any image-based skill is applied, including Box’s Image Intelligence Skill, Computer Vision will return even richer results to apply to the file in Box.

New speech services for audio analysis

Microsoft Azure announced several speech-related technologies such as the ability to define a set of vocabulary to recognize things like industry-specific terms.