US Television Advertising and AI

of two

AI in TV Advertising - Use Data

The use and presence of artificial intelligence (AI) in the television advertising industry is beginning to grow more popular in the U.S., as broadcasting networks such as NBC Universal (NBCU) and C-SPAN are using the technology for multiple purposes. NBCU specifically is using AI to scan TV show scripts while airing in order to match relevant ads to the TV content. C-SPAN has introduced AI technology to the network in order to automate transcribing and closed-caption processes with the intention of assistant viewers/readers in searching for relevant content. A study from TV Technology that was completed in Fall 2017 surveyed over 300 businesses in the media and enterprise industry. Results from this study showed that broadcast, cable, and governmental education TV networks have tested and are beginning to use AI technology to some extent. Below you will find a brief methodology section depicting how the information for this topic was gathered, in addition to two case studies on NBCU and C-SPAN and their use of AI for TV advertising, as well as the results of the study from TV Technology for an industry-wide use estimation.


In order to find content related to the use of AI in the TV advertising industry, multiple searches were conducted first for hard data on the topic. However, after multiple trials and error, no previously compiled reports or information from a single location were available. Statistics were more readily available for video-on-demand services (such as Netflix), and perhaps this may want to be addressed separately in another request. There were also many articles available that provided qualitative data on the use of AI in the TV advertising industry (such as for CBS), but only statistics for NBCU were officially stated. A search for more generalized data on the use of AI as a whole (as opposed to case studies for specific networks) produced the study by TV Technology. This report summarizes the overall use of AI in the market right now and even provides an idea of what percentage of companies in the TV broadcasting commerce are testing AI technology but not yet using it. As a result, this study will be consulted to gain a broader understanding of the use of AI technology in the TV advertising business as a whole.

Comcast NBC Universal

As NBCU increases the presence of AI on their network, the overall goal of the company is to meet consumer demands for fewer ads. This goal is ideally to be met by using AI technology to scan live TV scripts during airtime in order to match relevant ads to the show. In doing so, the network is expecting to decrease total advertising times by 10% on all of their channels, as well as decreasing the number of commercials aired during a break by 20%. These changes are expected to begin running live by the fourth quarter of 2018 whenever fall TV show seasons begin.

To build upon the use of AI to market more relevant ads, NBCU is also hoping to implement what they call the "60-second prime pod." This new advertising mechanism will be a result of the AI-chosen ads. The "prime pod" ads will run during the first and last commercial break of an aired TV episode and last for 60-seconds each. NBCU is expecting this new system to increase program lengths by 10 minutes, which in turn will allow the network to run 15 additional days worth of TV shows and other content in a year.

These changes and implementations as a result of the AI-scanning technology are expected to take place during at least 50 shows that are aired by NBCU on networks including NBC, E!, USA, and Bravo. The company also knows that they plan to invoke this method during Saturday Night Live on NBC, reducing the number of commercials by a total of 30% for this show alone.


In 2017, C-SPAN began using AI technology across their three TV stations to automate transcriptions and closed-captioning processes, as well as to automate image-recognition for their online content. At the moment, C-SPAN is using this technology to match speakers to a database of over 99,000 government personnel and transcription their words immediately. This technology is going to be saved into a system that will then allow users to search for content by words and have video content produced as a result. There are not any statistics on the success of this process yet, as it is still in the works.

TV Technology Study

According to the study completed in Fall 2017, approximately two-thirds of all respondents from the media enterprise industry have at least tested out and/or are currently using AI technology in some way. Of the types of groups and networks included in the study, the percents that reported testing/using AI are as follows:

— Production/post-production studios: 10%

The types of actions that are carried out through these networks and group with AI included:

In the broadcasting and streaming companies included in this study, some of the most popular uses of AI included indexing both current and archived data so that public users could immediately retrieve specific information through the AI system, as well as transcribing and translating all current and archived content. Further uses including monitoring content, advertising and placing content strategically.

It is interesting to note that the common actions noted in the study — metadata creation, clip distribution, automated captioning, and image recognition — are all actions that were noted to be used both by NBCU and C-SPAN, even though these networks were not specifically stated within the study. This leads to the conclusion that the information gathered by the TV Technology study could likely apply to both these networks and other major TV agencies in the U.S.


The use of AI technology in the TV advertising is likely to grow in coming years, as a large number of broadcasting and TV networks have reported at least testing out the technology, if not already using it in some form. NBCU and C-SPAN are prime examples of ways in which this technology can be used for advertising and expediting certain services in the TV advertising. Although not addressed in this report for being out of scope, another request may be of use on use data for AI technology in video-on-demand services, as there seems to be more hard data immediately available for this industry.
of two

AI in TV Advertising - Trends

AI in TV advertising is mainly being used to tailor advertising to specific viewers, in what is called programmatic TV advertising — which is both beneficial to advertisers and viewers, since it increases the efficiency of ads and reduces the number of commercials being broadcast.

Agencies have also been paying attention to AI in video advertising, though most have been focusing on online platforms — strategies that can be applied to streaming services, such as Netflix. Two strategies that have been experimented with, specific to video ads, are ads that change in real-time according to the specific viewer, and ads with flexible storytelling, which adjust stories to data.


In order to answer your request, we searched through reports and consultancies such as Nielsen, Adage, PSFK, IAB, GfK, and Ipsos — but none of them have published any studies listing hard data on AI usage in TV advertising. One reason for this is because this is a recent technology, and there is still not a lot of material to gather data from.

Therefore, we focused our research on qualitative data, case studies, and growing trends taken from specialized media and agencies, such as AdWeek, Divitel, VentureBeat, CognitiveScale and TV Rev.

Please note that due to the novelty of the technology and its application to TV advertising, the information and case studies are still scarce. In order to offer you the most data available, we expanded our research into online video advertising as well, which can be applied to streaming services such as Netflix and Hulu.



The main trend regarding AI usage in TV advertising is one that has already been successful in online advertising: programmatic advertising. Essentially, this means that TV programmers can tailor advertising to which viewers are on at any given time, according to data gathered through the TV.

As VentureBeat puts it, "your TV can learn about your habits in the way your web browser already does, allowing advertisers to present you with ads based on that information — so you’ll see fewer repetitive ads that you don’t care about. This means you and your neighbors may all be watching the premiere of The Walking Dead but seeing different ads based on your unique interests."

This has several advantages. For advertisers, it will maximize returns and cost-versus-benefit by slotting ads when they will have the most impact. For viewers, it will reduce repetitive and irrelevant ads, as they will only see commercials for those products that interest them.

Programmatic advertising will also allow advertisers to see which campaigns have the most impact and when. They can use machine learning to calculate the probability of success of a future campaign — which will, in turn, maximize returns even more.


NBC has started to experiment with a different type of AI-powered tool, matching ads to TV content. The station is linking commercials to specific shows, by scanning the content of the show and connecting it to relevant advertisers.

This more optimal placement of ads will ensure they reach suitable audiences, and the station hopes it will help reduce the total amount of ads by 20% — which will, in turn, benefit the audience since they will only see relevant ads in fewer numbers.


Several ad agencies have started working on ad strategies focused on AI-powered tools. Most are aimed towards online platforms, but some can be deployed in video streaming services, such as Netflix or Hulu.

One of these strategies, specific to video ads, are responsive ads that change course depending on who is watching them and what their interests are. These ads use data signals to point whether the viewer is someone who, say, would more likely travel to Asia instead of Europe or do so alone instead of with their family, as explained by AdWeek.

Another strategy is for ads to be built around "flexible storytelling," in which the specific story told in the ad will adapt to the viewer to make it more compelling and engaging. These ads will capitalize on data to build a story that really speaks to that viewer alone — personal to who they are.


To wrap up, AI in TV advertising is mainly being used to tailor advertising to specific viewers, in what is called programmatic TV advertising. Agencies have also been paying attention to AI in video advertising, though most have been focusing in online platforms — two strategies that have been experimented with have been ads that change in real-time according to the specific viewer, and ads with flexible storytelling, which adjust stories to data.