The vision for true next-generation communications involves a lot of computing power, a lot of analytics, applied to perhaps the ultimate form of unstructured Big Data: Real-time communications themselves. This translates into relatively new integration ideas around context--the idea that you could use the addressing and identifiers of realtime communications participants to pop up relevant documents and other stored data in near real time.
Another potential application is around analyzing the content of real-time communications. Contact center vendors have begun including speech analytics in their products, in order to give managers a more direct and potentially more accurate gauge of customer satisfaction than what can be extrapolated from metrics like first-call resolution.
As the "heaviest" real-time application, video is naturally lagging behind many analytics capabilities that focus on pure voice or text-based communications. And yet, as video becomes more popular in all its forms--real-time conferencing, store-and-forward, corporate broadcasting, simple ad hoc contact--more important and valuable information resides within those files; and as business-critical interactions happen over video (think Amazon Kindle Mayday), there are implications not just for revenues, but for governance as well.
I had a chance to talk last week with Ari Bixhorn, marketing VP at Panopto, which specializes in video search. Panopto started off as a research project at Carnegie Mellon University in 2002, where the technology found an early sweet spot: Recording lectures on campus for on-demand views. The technology was spun out as a startup firm in 2007.
Between 2007 and 2011, Panopto sold exclusively to universities, which continue to be its core market--the University of Arizona records more than 3,000 hours of video a week, and the UK's University of Essex averages 80,000 hours a year, according to Ari Bixhorn. However, the company's sales have broadened, and now include HR departments that use the technology for employee onboarding, to facilitate production and management of video guided tours of everything an incoming employee needs to learn. The platform was also adopted in 2010 by the New York Stock Exchange, which uses it for quarterly CEO updates.
Though the main focus started out as simply recording, the Panopto platform today is sold as an end-to-end package for recording, storing, streaming, and search; Bixhorn describes the platform as "A Sharepoint that's been built specifically for the needs of video."
The platform manages the entire video lifecycle, but video search is the most compelling element--potentially. Panopto actually relies on speech recognition with time stamps to enable its video search capability, which Ari Bixhorn said has reached an accuracy rate of 70-75%. Is that good enough? Ari Bixhorn said that for his early-adopter customers, it's a "huge benefit" to be able to search video using anything other than manually-entered metadata.
In addition, Panopto also uses optical character recognition to drive a text search function, so that if the video is of a presentation--whether a formal PowerPoint or an informal whiteboard talk--you can also search according to text characters; here the accuracy is 80%-90%, Bixhorn said.
In theory, video search ought to have about as many applications as video itself. In another type of use case, Bixhorn described a large electronics manufacturer that uses Panopto for a field service application: Techs can use their tablets to video a printer's model number and use that for quick access to documentation on that particular model.
The business case pretty much relies on the promised productivity gains, which seem at once obvious and difficult to quantify accurately. Panopto sells the software platform on a per-seat model. Anyone can record a video to the Panopto system via a client that runs on any device with a camera. The content management system is server-based and can be purchased either as CPE or hosted. It can integrate with Active Directory so that users are only allowed to manage and search videos that they have permissions to access in the first place.
If video ever does become the new voice, or even something close to that, enterprises will need systems that manage this vast store of content--for e-discovery, potentially; for internal purposes, and potentially to expose to other business applications--Ari Bixhorn said Panopto already integrates with Salesforce.com, SharePoint, and Microsoft Lync, among others, so it can serve as the back-end engine to video management and recording with these UC and collaboration applications.
The search technology may not be 100% yet, but as both hardware and software continue improving, it's likely to continue moving closer. Similarly, enterprises are just beginning to think about ways to integrate Big Data and analytics into communications. As video expands in the enterprise, the need to manage and leverage its content will likely increase.
Follow Eric Krapf and No Jitter on Twitter and Google+!
@nojitter@EricHKrapf
Eric Krapf on Google+
