US-based Chronicle Media has opened up access to Ubiquity, a platform that claims to identify the future audience for a video, series, or game before it's even released—including how many people will watch, in which countries, and what their interests are. It sounds like every creator's dream: picking the perfect topic and thumbnail before you even start shooting. Let's break down what the company is promising, what these promises are built on, and what can actually be verified today.
What is Ubiquity?
Chronicle Media is a US company founded by Aaron Sisto, Scott Greenberg, and Ollie Lewis. They unveiled their model on October 7 under the working name UBQT and called the platform built around it Ubiquity. On October 26, the industry publication Tubefilter reported that access was now open.
The company describes a three-step process the platform follows:
Audience Prediction
The model analyzes pre-publication materials—thumbnails, stills, audio, a short synopsis—and predicts which viewer groups will respond to it. According to the company, the output includes the audience size, geography, and niche interests: not just "males 18–34," but specific communities built around a genre, game, or artist.
Optimizing Thumbnails and Packaging
The same video can be packaged in different ways. The model evaluates various thumbnail and trailer options and suggests which one will perform best with the chosen audience segment.
Ad Buying
The third part is the automated launch of paid promotion targeting the identified groups, with continuous budget and settings optimization. This is more for studios and brands, not your average creator with a smartphone.
Where Does the Data Come From?
Sisto says the model isn't trained on AI guesswork but on the company's own proprietary data: Chronicle constantly runs advertising split tests across different viewer groups to see who actually watches what to the end. The company claims to work with over a million audience groups sourced from platforms like YouTube. The focus is on groups, not individuals—the company states it doesn't track specific viewers.
What Numbers Is the Company Citing?
All figures below are claims made by Chronicle Media itself; there is no independent verification for them.
| Claim | Figure |
|---|---|
| Audience prediction accuracy | 94–98%, but only "when fine-tuned for a specific franchise" |
| Private beta | 6 months, over 50 major clients |
| Audience groups in the model | Over one million |
| Podcast with a celebrity (a Fortune 10 client) | 150% audience growth in two months |
| Foreign-language series | 600% increase in daily revenue in the US over two weeks |
| New game | 1.5 million views and a 1.44-to-1 return on ad spend |
Separately, the company compared its performance to that of standard large language models (LLMs) that were asked to identify an audience for the same material. According to their data:
- 85% of the audiences identified by Chronicle generated significant impressions (150+), compared to 48.3% for standard AI models.
- Engagement was 46% versus approximately 25%.
- Full video completions were around 3,100 versus 1,800 for the best-performing LLM.
- The cost per action was 13% lower with twice the number of views.
What Can Be Verified?
For now, almost nothing—and that's the main reason not to drop everything and rush to get access.
The numbers were calculated by the company itself. Not a single independent measurement of accuracy has been published. The comparison with LLMs was also conducted by Chronicle—the test conditions and which specific models were used were not disclosed.
The 'franchise' caveat devalues the headline number. The 94–98% accuracy is claimed "when fine-tuned for a specific franchise," meaning when a property already has existing installments, viewers, and a watch history. This is exactly what a creator planning to shoot their first video in a new niche doesn't have.
The case studies aren't about individual creators. A podcast from a major corporation, a foreign series, and a game with an ad budget—these are clients with studio-level money. The company hasn't shared how the model would perform on a channel with just a few thousand subscribers.
There's no price tag. The Chronicle website has no pricing tiers or a "try now" button. There's an application form asking for your name, email, company, and social media links, after which they will "be in touch."
What Should Creators and Brands Do Now?
You probably won't get access to Ubiquity anytime soon. But the core idea—testing your packaging before you invest heavily—works even with free tools.
For short-form video creators. Thumbnail testing is already built into YouTube Studio: you can upload up to three options, and the platform will show them to roughly equal-sized audience groups. After a few days (sometimes up to two weeks), it will declare a winner based on watch time share. It's basically a mini-Ubiquity for zero dollars. Another trick is to shoot a cheap, short version of an idea before investing in a big production; the retention on a draft will tell you more about the topic than any prediction.
For brands. The logic is the same: start with several low-cost versions of a video and thumbnail on a small budget, then scale what works. We broke down which UGC tactics have the longest shelf life in our article Which UGC Videos Have a Longer Shelf Life: An Analysis of 8,107 Meta Ads.
Both creators and brands should remember one simple thing: audience prediction is not a guarantee of views. It reduces the number of shots in the dark, but you still have to do the shooting and editing.
What's Still Unknown
Chronicle hasn't named a price, specified who gets access or how quickly, disclosed the methodology used to calculate its accuracy, or clarified if the platform works with Russian-speaking audiences. At the time of this article's publication, no independent reviews of the model have been published.
Frequently Asked Questions
What is Chronicle Ubiquity? It's a platform from the US company Chronicle Media, based on its UBQT model. It promises to predict the future audience of a video, game, or series before publication and help optimize its thumbnail and packaging.
Can a neural network predict how many views a video will get? No one can give you an exact number. Chronicle claims 94–98% accuracy, but with the caveat that the model must be fine-tuned for a specific franchise with a known audience. For a new channel in a new niche, that data doesn't exist.
How much does Ubiquity cost? The company does not publish its pricing. Access is granted via an application on their website.
How can I gauge my audience before releasing a video for free? By using the thumbnail test in YouTube Studio, creating a short draft of your topic instead of a full-blown video, and analyzing how the same topic performed on similar channels. It takes longer than clicking a button, but it's available today.
How is this different from regular analytics? Analytics show you what has already happened with a published video. Chronicle promises an evaluation before publication—based on the thumbnail, stills, and audio. How accurate it is remains to be seen.
We've written about how creators are already using AI in their video workflow in our breakdown of YouTube's report on inserting yourself into a generated scene, and about the adoption of neural networks in our piece on How Many People Use Neural Networks in Russia.
While some promise to predict your audience, what's already making creators money is a simple concept: a paid brief for a video. On Prime Oracles, brands post these tasks in the "Content Rewards" section. A creator can accept a suitable brief with one click and get paid for the views their video generates.
