AI search engine Perplexity has begun sharing advertising revenue with news organisations and hired its first head of publisher partnerships to nurture those relationships.
Perplexity allows users to ask questions, including about current news events, and receive AI-generated answer summaries as well as a list of several sources, pictures and videos. It claims to be receiving about 20 million queries per day, up from about 2.5 million at the start of 2024.
It differs from OpenAI’s ChatGPT because it does not build (and therefore need to train) its own foundation models – instead it answers queries using a web index with information about the content on various websites.
Perplexity first revealed plans to share revenue made from advertising in search results with the publishers whose content is cited in response to a query in July.
The ads, which began appearing in November, take the form of a sponsored query in the “related questions” section under Perplexity’s AI-generated answers. After they are rolled out fully, the goal is that ads will appear against “most, if not all, of our related questions across different categories”, head of publisher partnerships Jessica Chan told Press Gazette.
“That obviously will just mean more opportunity for publishers to make money as well.”
Small amounts have now begun being paid out to publishers. Chan added: “It’s going really well. We’re in a charter period right now, and so we are testing to ensure things like brand safety. And then once we’ve got the learnings, we expect to scale that up more aggressively by next year.”
The percentage of revenue shared with publishers in the programme varies depending on the number of links cited and it is believed to be capped at 25%.
As well as revenue sharing, the publisher programme gives partners access to Perplexity’s APIs [application programming interfaces] so they can create their own custom answer engines for their websites, and one year’s free access to the Enterprise Pro corporate package for all their employees.
Publishers who signed up to the programme at its launch were: Time, Der Spiegel, Fortune, Entrepreneur, The Texas Tribune and WordPress owner Automattic. They were joined this month by more than a dozen others including Adweek, The Independent, US local publishing owner Lee Enterprises, Los Angeles Times and World History Encyclopedia.
Chan, who previously led publisher partnerships at Linkedin for more than eight years, joined as Perplexity’s first head of publisher partnerships in September.
Perplexity created the new position, according to Chan, because it was “pleasantly overwhelmed by the demand” for the publisher programme.
“We have received more than 100 inbound messages from publishers and it just became very clear that we really needed someone full time to lead this programme,” she said.
“And so for me, my role really involves expanding this programme to bring in new publishers to the programme, of course support the existing publishers, and then continue to establish… new types of benefits that we can offer partners,” she added, suggesting more extensive analytics tools and discoverability on its Discover page as examples of something that might be “mutually beneficial”.
In the summer both Forbes and Conde Nast sent legal letters to Perplexity over accusations of their content being used without permission. News Corp subsidiaries Dow Jones and NYP Holdings, the publishers of the Wall Street Journal and New York Post, have since launched legal action.
However, Chan said Perplexity continues to welcome conversations with all publishers, even those that have previously raised concerns about its model.
And she said Perplexity obeys sites that have disallowed robots.txt crawling despite some of the allegations that have been made against it including from plagiarism detector Copyleaks last month which claimed the AI start-up was paraphrasing “substantial portions” of paywalled articles.
More publishers are unlocking robots.txt as they sign up to the partnership programme, Chan added.
“With certain partners, it’s really been fascinating,” she continued. “A lot of the partners have been eager to want to test different variations of their content. So whether it just be public content or publicly available content, or content behind a paywall, or sometimes even archive content, the publishers have been eager to test a bunch of different models.”
Perplexity working with publishers who ‘want to be pioneers’ on AI
Chan said that since she joined she has “seen a lot of excitement from publishers across the globe… wanting to be part of something new.
“There’s a lot of learning on both ends, I think. But I think for a lot of these publishers, they want to be pioneers in the conversations that we’re having and figure out how we want to collaborate with them.
“And so, like I said, it’s been, it’s been incredibly positive from from the experience that I’ve had thus far, and several dozen conversations I’ve had with partners around the globe.”
Chan went on to say that a lot of publishers “are in the learning phase right now, really trying to understand how AI works and how Perplexity is differentiated as a platform…
“We designed this programme to basically support publishers in the long term, and so for us, we are really trying to work collaboratively to figure out what is the right sort of model to work with partners long term to align incentives.
“And so I think once they start to understand that, and once they understand that we are fundamentally very different, we’re not trying to license a bunch of data to train any models, I think that’s where they start to appreciate the programme a lot more.”
At Linkedin, Chan worked with both publishers and other content creators to bring their content into the platform’s ecosystem.
She said she brought a number of lessons from Linkedin to Perplexity because there she was “constantly trying to find much more sustainable, collaborative ways of working with publisher partners, as opposed to some competitors”.
Without naming them, she said some platforms were “a little bit more pay-to-play, but we were really trying to seek out sustainable models that aligned incentives, both potentially from a financial as well as from a product collaboration standpoint.
“I find that is very much the same thing here, which is: we want to be friends to the publishers and so how do we do that in the most mutually collaborative way?”
Chan added that her time at Linkedin, from 2013 to 2021, coincided with media companies thinking about how to evolve and optimise for social media – and that she is now at a similar forefront with the AI evolution.
Email pged@pressgazette.co.uk to point out mistakes, provide story tips or send in a letter for publication on our "Letters Page" blog
