Annelies Jansen, Chief Business Officer at ProRata.ai, leads business strategy at the Los Angeles-based AI company founded by Bill Gross. ProRata creates fair AI-publisher partnerships through transparent attribution technology and 50/50 revenue sharing, working with major media partners including The Financial Times and The Atlantic to ensure content creators receive compensation when AI systems use their work. In this interview with Jansen, we find out why publishers are drawn to Prorata’s technology.
What is Prorata.ai?
We’re helping publishers to build new business opportunities using AI. We’re an AI monetisation engine for all content creators. What does that mean? We want to make sure that all content creators are being compensated and credited for their content use by AI. We don’t take content from publishers without them knowing as we have licensing agreements with all the publishers we work with.
What are your products?
This license model allowed us to build an attribution technology and this model is the foundation of our business. The attribution model enables each GenAI answer we generate to bring a clear attribution to what content sources were used and at what percentage. This means we can compensate all content creators based on their contribution.
We’ve created an ethical answer engine called Gist AI that’s available now as an embeddable widget that can be integrated within any of our partner’s media channels.
Finally, we have our ad product which lets publishers make money off of ads shown in the GenAI answers that show on their page.
Tell me about the attribution model.
50 percent of citations from AI models are hallucinated – be careful what you assume when you see a citation attached to an answer you get. In essence, our attribution model is based on the licensed content relationship we have. We’ve built a proprietary tech that dissects the content when it’s ingested into our database. It dissects it up to the original factual claims level using neurolinguistic engineering with vector-based fingerprinting.
It’s not attributed just at an article-level, but an original factual claims level.
What is your business model?
50 percent of revenue goes to Prorata and 50 percent goes to the publisher. Publishers can make money in their own media channel and if their content is attributed on a different media channel.
For example, Gist Answers is embedded in Women’s Health, we just launched with them. When a user asks a question, an answer can be generated from content from any of our partners. So those other publications can generate revenue from an answer that’s not happening in their own media channel.
If I’m the Sacramento Bee, for example, and I have Gist Answers integrated in my media channel, and for a particular answer it uses The Guardian’s content, then The Guardian will receive from that 50 percent prorata from their contribution to that answer.
Our goal is that our attribution bar will become an accepted product within the industry.
The other revenue stream for publishers is when publishers monetise using their own ad slots to be served alongside our answers. Our ad product is served within the answer.
Who uses Prorata?
It’s for every kind of content creator. Our user is anyone from the large traditional titles, to the smaller creators building a million followers. We have content guidelines to make sure these publishers are ones that value journalism and deliver original content and not synthetic content, in-sync with regulatory and publishing associations, and IAB standards.
What’s your mission?
We wanted to show publishers that when we use their licensed content, we can build quality answers that come with attribution.
We’re a non-exclusive product solution. This isn’t about the industry having to collaborate around a negotiation table. This is a product solution that’s non-exclusive. When publishers leave, they take their content with them. We don’t keep that content for future-answers. The time is now. Consumers are adopting AI fast and it’s only going to get faster. We made this solution easy for any publisher to use – all it takes is licensing your content and four lines of code.
What pain points does Prorata solve for publishers?
When publishers join Prorata, they also join our movement. For publishers, content is taken without credit or compensation, nor analytics. Publishers are losing control over where their content appears. For those models that do attribute, there’s inaccurate, hallucinated outputs with wrong attributions, or no attribution at all.
This can jeopardise a publication’s brand reputation because you don’t want your content to be used in a hallucinated answer. The zero-click search environment is a real threat to the industry. Google’s AI Overviews – which appears at the start of a search – don’t deliver links back to the original content source.
The relationship between publishers and AI is broken and you see that in the poor traffic numbers today with all of our partners. That traffic loss needs to be compensated, so the opportunity we bring to publishers with our embeddable widget is that you can create more engagement on your media in-line with where user behaviour and appetite is moving to today – a user is expecting to have an AI companion experience wherever they are.
Does it bring publishers closer to their audience?
Yes. It helps research and development because teams can find out what content is most used as answers to their audience based on the questions they’re asking. You can get the insight of what prompts are being generated on your website. We share analytics with publishers to help them optimise and prepare for the future.
Is there a difference between the quality of output from an ethical AI model over a more traditional model?
We have a team looking at the answer quality of our outputs, comparing our answers to those coming from LLMs. You don’t need billions of articles to generate a good answer, you only need millions. Our database right now is approaching a hundred million articles. I claim this is the largest database of licensed content for GenAI answers.
What might incentivise users beyond ‘ethics’?
With the ‘ethical’ message, you aren’t going to convince the average consumer. The average user wants convenience. We are working with publishers to make sure the experience that takes place within their media channel, is optimised and delightful for the user where they don’t feel compelled to move away from their AI experience within the media channel.
Users are already invested in AI, so the idea is that if you bring the AI into the media channel this will be a good acquisition and retention tool because it enhances the user experience.
The other benefit is that the AI search advertising market is predicted to be the fastest growing ad market we will have ever experienced. If you bring AI engagement into your media channel, there is also the opportunity for there to be an AI search ad shown within the answer – that enables you to build AI search advertising in your own media channel, instead of having to depend on receiving ad revenue from AI companies, and having to negotiate on a platform you can’t control.
Why do you think the AI search advertising market will be the most lucrative? And what happens over the next two years to make that happen?
We see this market growing because of fast AI adoption from users. When it comes to the user, advertisers follow user behaviour and want to be where they are. That’s why this AI search ad market is so interesting, because it’s not just using traditional ad formats within the new environment. This opens up the opportunity to bring new ad solutions to users using AI.
Our ad tech is bespoke, and effectively uses AI to have a better understanding of the prompt, the answer, and the advertiser’s content. This enables the presenting of a contextual ad that doesn’t use cookies or first-party data to serve you something good.
