Microsoft is looking to claim itself a spot beside the top dogs competing on the video and cloud front.
It says data is the differentiator.
“Unlike other providers who use your data to sell groceries or other services that compete [with your business], your data belongs to you,” Martin Wahl, principal program manager for Microsoft, told a room full of media and entertainment execs Tuesday at an AI Meets Media innovation summit.
Wahl, who works with global customers of Microsoft’s Azure Media Services group, seemed to be taking a swipe at the competition, namely Amazon and IBM.
Both have built and acquired their way into cloud video stacks. Amazon bought back-end video platform Elemental Technologies and merged it with its cloud infrastructure Amazon Web Services. IBM rolled its purchases of video service providers Clearleap and Ustream into IBM Cloud Video.
“Microsoft for years has looked at how we take cloud artificial intelligence and apply it to video,” Wahl said. “My team is responsible for making what Microsoft is doing in cloud AI … applicable in places like media and advertising.”
One ongoing headache for media and entertainment shops is wrangling their ever-growing volume of video streams on multiple channels, devices and operating systems.
Internal production teams want to reduce latency on content uploads and speed up video delivery, while advertisers like Coca-Cola and broadcasters like Sinclair need a better way to tag video content to improve discovery and personalization.
A combination of cloud and AI promises to help solve these pain points. Microsoft is doubling down on both.
Microsoft already built cloud-based technology that recognizes facial movement, along with audio detection and sentiment analysis to determine whether someone is happy or sad.
More recently, it rolled out a cognitive video service called Video Indexer. Historically, companies were forced to manually comb through and tag objects in video streams, but artificial intelligence should help expedite that process.
“Video Indexer lets you access all of your video metadata and correlate it,” Wahl said.
By moving “object tagging” to the cloud and applying machine learning, an advertiser, agency or media company can analyze what a video stream consists of frame by frame.
For instance, a CPG brand might want to see the top 10 occurrences of a particular product within commercials on-air or in digital, or an advertiser might want to swap out creative if a certain actor appears in a show in order to improve the contextual relevance of an ad.
Cognitive also comes into play by determining the tone and sentiment of people in a video frame using speech and keyword identifiers, which media companies can tap to tailor content more effectively.
Microsoft’s Video Indexer tool provides most of the necessary “raw” back-end support, such as video transcoding, hosting, storage and delivery. But it relies on partners, including video platform Ooyala, to provide the “practical” use cases such as, for example, improving revenue or return on ad spend.
Taken together, “this data can determine much more than, ‘These users watched this genre of show,’” Wahl said.
“It’s a matter of taking your existing metadata, re-indexing it and showing you new insights with cognitive,” he said. “We’re giving you all of this metadata, [analyzed through AI], which can be used for production, editorial, advertising.”
