Breaking the Mold on Brand Suitability with Contextual AI: SeekrAlign H1 Retro

Davis Cannon
Product Manager
June 18, 2024
breaking the mold on sustainability with contextual ai: seekralign h1 retro
Product

Bringing clarity to your campaigns with transparent AI

In December 2023, we launched SeekrAlign to solve a serious and growing need for advertisers: evaluating podcast brand suitability accurately and at scale. SeekrAlign made it possible for brands to evaluate millions of minutes of podcasts and understand the true spirit of the conversation, empowering them to make more informed decisions about which podcasts to sponsor and grow their reach responsibly.

Since its initial launch, our team has continued to expand SeekrAlign to serve advertiser and publisher needs across the industry, with breakthroughs in audio, news, and programmatic. At the same time, our AI model has continued to learn from user feedback and increased data inputs to analyze content with even more precision and accuracy. We’ve built more explainability features, more transparent AI scoring, and more data insights into SeekrAlign to make content evaluation as simple, clear, and accurate as possible.

Keep reading for a recap on the tools leading marketing teams like SimpliSafe, Babbel, and Indeed use to simplify their brand suitability process, and explore what’s next on our roadmap.

What’s new in SeekrAlign

AI explainability enhancements

Advertisers need to understand the true suitability risks in podcasts to make informed media decisions. We’ve released new score snapshots on the show and episode pages to give users a clearer view into a show’s Civility and GARM risk levels. To dive deeper, users can click on a risk in the snapshot and review flagged content directly in the transcript. These new data visualizations and explainability features make brand suitability measurement faster and clearer than ever.

GARM risks

More model transparency

Our users wanted to understand the reasoning the AI used to determine show scores. Quick tooltips on the show and episode pages give advertisers the context they need, when they need it to speed up their evaluation process. We added full documentation to the SeekrAlign dashboard to explain the principles and logic behind each evaluation. We’ve also included Civility Score and GARM definitions and examples, giving more transparency and explainability into our AI model’s decision-making process.

Podcast data API access

SeekrAlign APIs are now available for news and podcasts. This gives businesses the ability to adapt our scoring data to their specific needs with more speed and scale through customized reporting and platform integrations.

Interested in integrating with Seekr? Contact [email protected] to inquire about API access.

Upload and score your own podcast list

Brands want to make sure their podcast targeting lists are brand suitable. We made it easy for users to score any podcast on the web, even if it isn’t currently available in our index. They’ll simply upload their custom shows list and SeekrAlign will deliver the results. This empowers users to discover new shows and expand their reach while ensuring they’re aligning with shows that fit their values.

More precise AI scoring models

Advertisers rely on the accuracy of our suitability scores to make informed media decisions. To ensure the most precise results, our team makes continuous and ongoing improvements to the SeekrAlign model through human-in-the-loop training, where we collect data from user feedback and internal reviews to retrain and fine-tune the AI model. And through our collaboration with Intel, we’ve optimized inferencing and increased the amount of data we can process to enhance the model faster.

In our latest model release, we improved F1 scores (precision and recall) for all scoring detectors. This means sensitivity enhancements for Civility Score and GARM detection in episode transcripts and refined podcast show scores, ensuring each line of text is scored accurately and consistently every time.

As SeekrAlign grows, we will continue to enhance the model based on user feedback and data collection. The more data we receive and the more content the AI scores, the more accurate and reliable it becomes.

Over 62 years of audio scored

Brands want to reach more suitable audiences across the top 50 and the longtail of podcasts. In the last quarter, we tripled our podcast index from 10,000 shows to 42,000+ shows, and that number continues to grow. Users can access scores for top shows like The Joe Rogan Experience, The Ben Shapiro Show, and Call Her Daddy, while discovering more niche shows that align with their audiences’ interests.

seekralign brand products

Breaking into news scoring with SSP partners

The brand suitability challenge doesn’t end at podcasts. Brands are looking for more transparent and contextual brand suitability scores across platforms. With this understanding, we integrated our content evaluation technology into supply side platform (SSP) Equativ to score web-based news content at scale.

This partnership enables Equativ’s advertising clients to evaluate news content at the URL level based on journalistic principles including subjectivity, clickbait, and political lean.

Try out the new SeekrAlign features yourself

Book a Demo

What’s next on the roadmap: content evaluation for your entire digital advertising plan

The advertising industry needs AI to help them scale campaigns responsibly across platforms. In our upcoming releases, we’ll enhance SeekrAlign’s dashboard to make evaluating podcasts simpler and expand our AI model to detect more risk signals across channels and mediums.

Political lean scores for podcasts, news, and the web

Understanding the political lean of content helps advertisers have more control over where and how their brand is perceived. Soon, our political lean indicator will be available for podcasts, news, and web content to reveal which stance a piece of content takes on the political spectrum—whether that’s left, right, center, or somewhere in between.

To reduce bias and overgeneralization, our AI analyzes the exact content of an individual episode or article, rather than just the overall publisher. This allows brands to confidently reach their audiences in places they may have previously avoided.

seekr political lean AI content scoring for news web and audio

GARM scores for web content

Advertisers need to detect brand suitability risks across the web. That’s why we’re expanding our programmatic capabilities to score all web content for GARM risks. Offered through integrations with Equativ and our new programmatic partners PubMatic and OpenX, brands will be able to unlock more reach in these curated, quality environments.

Building the omnichannel solution for content evaluation

In the months ahead, we plan to expand SeekrAlign to video, social, and a broader section of the web. We’ll also train our AI model to score content for IAB taxonomy and other framework criteria for more context and clarity into brand suitability. As our AI capabilities grow exponentially, SeekrAlign will be able to rate and score any piece of content on the web.

Where will we go next?

We’re on a mission to make SeekrAlign the omnichannel solution for content evaluation and brand suitability. Since launching our podcast solution in December 2023, we’ve grown Align to serve customer needs across platforms, industries, and use cases through our proprietary AI models.

As our AI continues to learn, it will be able to train itself to evaluate content for any scoring framework across the entire web, providing insights and information back to users so they can make more informed media decisions.

Bring more clarity to your campaigns with SeekrAlign

Book a Demo

Get the latest Seekr product updates and insights

This field is for validation purposes and should be left unchanged.