Broadpeak also says the solution provides “full visibility into the advertising platform, offering enhanced, data-based observability and robust security.”

The company says broadpeak.io DAI as a Service streamlines the delivery of targeted ads in live programmatic workflows and unknown ad pod sequences, while enabling publishers to “easily perform branding with slate or promo content.” A live pre-roll feature allows operators to insert an ad pod when viewers start watching a channel or change channels. According to Broadpeak, this creates additional monetisation opportunities. broadpeak.io DAI as a Service also supports pre-roll, mid-roll and post-roll targeted ad insertion for VOD and catch-up TV.

Mathias Guille, Vice President, Cloud Platform at Broadpeak, said: “Content monetisation is critical to video streaming success, and we’re excited to provide customers with a one-stop-shop solution that simplifies targeted advertising. broadpeak.io is the first self-serve software solution to offer Dynamic Ad Insertion as a Service that is agnostic to the different components of the video delivery chain, reducing the complexity typically associated with end-to-end advertising ecosystems and speeding up the monetisation of advanced OTT services.”

Broadpeak will demonstrate broadpeak.io DAI as a Service at IBC2022.

The rapid growth of streaming TV consumption makes it imperative that networks, studios and MVPDs have a ‘grown up’ connected TV offering – one that allows them to compete for every eyeball and ad dollar that migrates into the premium digital video environment. This means:

• Holding onto audiences, and attracting new viewers, by ensuring a consumer-friendly ad experience with modest ad loads, frequency capping and relevant ads.
• Enabling ad insertion (and targeting) at huge scale with total reliability, so that mega-events including sport and breaking news can be fully monetized.
• Maximizing the value of inventory by carefully managing direct sales with multiple programmatic demand sources.
• Effectively using audience data to achieve better targeting and a personalized ad viewing experience, even in a cookie-less world.

The pandemic accelerated all digital behaviours, including TV viewing. Mass-scale connected TV was tomorrow’s opportunity and challenge, but tomorrow arrived early. Join us live to hear how networks, studios and distributors can rethink the ad-supported TV model and ready themselves for connected TV primetime.

Register here.

This webcast explores:

• What a holistic ad performance and revenue strategy looks like, and how you find the sweet spot where ‘audience + avails = best possible yield outcome’, and how you avoid the one extra ad that loses you money, as viewers ‘tune out’.
• The potential for greater integration and synergies between SSAI and audience data and targeting capabilities –and how this could be implemented.
• How to manage ad frequency in a multi ad server (direct, programmatic, affiliate) environment and so ensure ad repetition does not degrade the ‘premium’ TV experience.
• The next steps in dynamic ad insertion, including the benefits of server-side ad insertion and client-side ad insertion, and how these can be used successfully in a hybrid model.
• The potential for a multi-DAI vendor strategy (similar to multi-CDN strategies) and how this could influence ad delivery performance and costs.
• How to manage multiple inventory owners within the same video session using an understanding of who has the ‘right to sell’, plus the ad routing and decisioning processes that support sales business rules.
• The special considerations when streaming ad-supported content into MVPD set-top boxes.
• How to monitor and diagnose the causes of ad errors, and so increase ad delivery reliability.
• How organizations are managing ad insertion for very large scale events, including the implications for ad insertion infrastructure

VideoAI was released for general use by Comcast Technology Solutions (CTS) in January. It is a fully managed, 24/7 software-as-a-service offering, so media companies can pay for the applications on a per minute basis. According to Bart Spriester, VP & GM of Content & Streaming Providers Suite at CTS, the key differentiating point for VideoAI is the provision of pre-trained algorithms which mean the applications work out-of-the-box. Other vendors provide the AI detectors but leave media owners to work out their own algorithms, he claims. The VideoAI algorithms draw upon millions of hours of AI learning at Sky, Comcast and NBCUniversal.

VideoAI is also fully automated – there is no manual work involved when segmenting show introductions or credits once they have been detected, for example. “AI technology can be focused not only on enhancing the end consumer experience but on reducing the operational cost for content providers and operators,” Spriester points out.

The immediate applications available with VideoAI are:

• The solution can be used to generate specific metadata to support content chaptering with automated titles and summaries, plus smart thumbnails. Intros and credit rolls are detected to support this. VideoAI can also help detect ad breaks to support dynamic ad insertion.
• Contextual advertising. The content is analysed for brand opportunities and for advertiser sensitivities. CTS says this capability improves the value of every ad break.
• Live-to-VOD. Thumbnails, titles and summaries are created on-the-the-fly as a live event airs, ready for fast on-demand availability. Highlight reels can be automatically created during live sports events – and these are shown on-screen to enhance the sports viewing experience.
• QoE checks. Video assets can be automatically analysed to identify and tag key onscreen moments such as hard cuts, black frames and transitions, or audio silences or specific sounds. Alarms can be set to warn operations teams if something is wrong. If there is silence, for example, is there supposed to be silence in the programming or has the stream failed?

Comcast Technology Solutions says content owners, operators and even advertisers can work with the company to develop their own business use-cases. And Spriester promises that even small customers can influence the CTS development roadmap for the new product.

In terms of workflow, the VideoAI detectors sit downstream of audio and video encoding/decoding and closed captions creation. Customers with content in the cloud (and Spriester says this would be most of them) can provide low-resolution proxy assets for the VideoAI detectors to check, prior to the AI-based applications being performed. Enriched metadata is then sent back. It would also be possible to run detectors downstream of full video files, and it is possible to analyse live video as long as CTS can site its detectors in the cloud where the live signal comes from (proxies cannot be used with live).

CTS is in talks with several potential customers outside the Comcast group, with segmentation for auto-chaptering one of the most sought-after applications. Spriester emphasises: “This is not just another AI tool. We are bringing the pre-trained algorithms on a SaaS-based model. Our solutions have been proven and deployed at scale at some of the largest operators and content creators in the world.”

Fraser Stirling, Chief Product Officer at Sky, adds: “VideoAI from Comcast Technology Solutions is a potential game-changer. With VideoAI as a managed service, companies can quickly launch a range of AI-powered video services at scale, and benefit from investments we’ve made internally, to drive their own business objectives forward.”