Major Round of Funding for the Development of a Video AI Architecture

We recently announced a major new round of funding to enhance our platform and services, as well expand our global sales and marketing efforts. This is a major vote of confidence by our new and existing set of investors in both our progress-to-date with over 200 international OTT service providers and publishers as well as our vision for the future. Conviva has been focused on powering the best possible consumer experience of premium streaming video on the internet for over ten years. From that foundation, we have built an unprecedented customer base with global data that has defined and refined our machine learning algorithms to the level of maturity we think is unmatched.

With these new funds, we plan to expand and enhance these algorithms that are at the core of Conviva’s platform – we call this our Video AI Platform. The two artificial intelligence-based components of the platform are an OTT Video Graph and Video AI Models. Before going more into what these components do and how they work, let’s first explore why the streaming video industry needs AI…

We have seen explosive growth in viewing hours, peak concurrent plays, devices, and the content that is available across all OTT services. While bandwidth continues to increase as well, the reality is that consumer demand for watching video on the internet is far out-pacing the internet’s ability to deliver it from the core to the consumer device. Furthermore, the internet was not architected to support such video transmissions. This has been one of the primary factors driving the growth of the CDN market as well.

OTT publishers and pay TV operators need help to overcome this gap between consumer demand and internet capacity – this is where AI can help. The most successful form of AI today is machine learning and at the heart of machine learning is an insatiable appetite for data. Machine learning is an AI discipline that uses data to define algorithms – in essence, it uses huge streams of data to write the code that will help understand what is happening at any moment or predict what might happen next. The many billions of hours of video viewing that we see is the perfect fit to develop or train machine learning. Furthermore, our customers can greatly benefit from any help with understanding the state of their OTT service delivery as well as help with predicting when things might be going wrong. For example, wouldn’t it be great if your favorite OTT service could know instantly that your streaming experience was getting worse, understand why, and get prescriptive advice on how to correct it within seconds? If you are watching your favorite show or sports broadcast and you experience buffering, you want it gone immediately or you are likely to stop watching.

As OTT audiences reach traditional TV-scale, there is simply no way for OTT providers to build a team to monitor and analyze the performance of every single one of their viewers – yet consumers demand that they do. This is the case for AI – automated systems that can learn and adapt to the dynamic nature of both the internet and our viewing patterns. The Conviva Video AI Platform assists our publisher customers with detecting problems whenever and wherever they happen, diagnosing the root cause of the problem, and finally offering predictive analytics to head off problems before viewers see them. These are the three classes of AI Models we have deployed:

1. Detection
2. Diagnosis
3. Prediction

Conviva’s Video Graph powers all three of these models with a dynamic representation between all the applications, channels, and shows we are all watching intersected with all the devices, CDNs, and ISPs we are connected with to deliver the streams. Much like a social graph captures relationships between people and their interests, the Video Graph captures the relationships between video content and the infrastructure used to deliver and consume it. This graph is a form of contextual memory for our Video AI Models helping them understand the relationships and therefore do incredibly fast diagnosis and prediction.

With this latest round of funding Conviva will be helping customers more intelligently and effectively deliver the next generation of TV. Exciting times!