In a new publishing deal, Causaly and Sage have announced a partnership to bring full-text scientific peer-reviewed literature into the world of AI-powered drug discovery research.
According to Sage vice president Katie Metzler, who leads the publisher’s global licensing team, this deal follows a similar partnership that Sage struck in January 2026 with Consensus, an AI workspace for scientific research based in San Francisco. The deal also marks Causaly’s first partnership with a major STM publisher.
“It’s an example of a broader shift happening with AI agents reading the full text of journals and providing human researchers with the sections they need, showing them how ideas are connected via their knowledge graph, saving time over traditional search and discovery methods and driving new insights that they may never have found via a traditional lit review methods.”
Headquartered in London, Causaly’s mission is “to accelerate discovery in life sciences through transformative AI technologies,” offering new ways “to find, visualize and interpret biomedical knowledge and automate critical research workflows.”
The newly announced partnership grants Causaly AI agents access to digest the full text of peer-reviewed papers from a group of some 400 Sage journals. Relevant insights from those articles are served directly within the Causaly platform, helping mutual customers “unlock deeper evidence” from their existing institutional licenses.
“Sage has always been committed to connecting researchers with knowledge in the most useful and meaningful way,” says Bob Howard, executive vice president, global journals at Sage. “With Sage content deeply integrated inside Causaly’s agentic AI platform, we can extend our reach and impact where researchers do their important work. Access and intelligence go hand in hand, and this partnership puts them in one place.”
“Partnering with Sage gives our customers something they have been asking for: a single platform that brings together evidence and governed scientific reasoning,” said Marco Costa, COO at Causaly, in a press release. “When our agents read the full paper first and surface the most relevant evidence and insights right in scientists’ workflows, our customers can research with more confidence and get answers faster.”
Causaly selected some 400 journals from Sage’s portfolio of more than 1,500 research journals. While Sage’s traditional strength is in social sciences, it buttressed its biomedical research strength with the 2025 acquisition of Mary Ann Liebert—the founding publisher of GEN. Liebert published more than 100 peer-reviewed titles, many of which are included in the new partnership.
“Causaly’s powerful AI platform, combined with trusted scientific content from Sage and Mary Ann Liebert journals, enables researchers to uncover insights faster and make more informed decisions,” Howard told GEN. “The future of drug discovery will be shaped by partnerships that combine authoritative scientific knowledge with domain-specific, agentic AI. Together, we are helping to make that future a reality, delivering greater value to our shared pharmaceutical customers and accelerating the pace of scientific discovery.”
Drawing upon the full text of Sage journal articles, including the methods, results, tables, and supplemental data, Causaly’s AI agents surface a comprehensive picture of the study.
The Sage journal full-text integration is now available to all Causaly customers as a separate add-on to their existing platform subscription. Sage subscribers can link straight from Causaly’s extracted evidence view to the full article on the Sage website. Causaly customers without a Sage subscription can view a snapshot of the full-text article, along with an in-platform pathway to purchase the article on Sage’s platform. By running its full-text relevance analysis before any link-out, Causaly customers can assess the value of a given paper before reading or purchasing it.
RAGs, not training
Metzler manages Sage’s global licensing team that was involved in negotiating this and other retrieval augmented generation (RAG) licensing deals. She gave GEN some background on the partnership and its broader significance.
It is important to distinguish between AI licensing for training and RAG deals, Metzler says. “There are two kinds of AI licensing. There’s licensing for training, where the content is used to train the underlying model. Then there’s licensing for RAG, which does not allow training of the underlying foundation model but instead allows the licenser to create a vector database of the content—embeddings of our content—then the agents retrieve snippets of the content to display to users.”
RAG deals with other companies are in negotiation, Metzler says. “From the Sage perspective, this is a story about how discovery is changing, how the behavior of researchers as a result of AI is changing. Increasingly, the starting place for a researcher’s journey is not on a journal platform or a Google search page, but instead on some kind of AI-powered natural language tool. Our strategy around discovery needs to evolve, so licensing is just one part of that.”
Like most other publishers, Metzler says that Sage is assessing how AI is changing different parts of the discovery process—how researchers find content, how they access it, ensuring that that access is rights compliant and that publishers are protecting the rights to published content. “There is also the trust piece,” she said. “How does AI change in a discovery context, how do you know what to trust when you find it through an AI-mediated platform?”
Publishers also have to consider the measurement of that usage. “How does that look different than it did in a world where we were thinking about organic web search and library discovery services? Those channels still exist, but there’s now a number of mediators that are changing that picture.”
Spirit of learning
Like Sage, Metzler said Causaly is entering this partnership in the spirit of learning. “We are both learning together about what the impact is on the value that they’re able to offer their customers and also the referrals or usage that drives to our content. It feels like this is a partnership, not just a licensing arrangement. I think there is a lot to learn on both sides.”
“We’re still as an industry figuring out what good looks like, in terms of these answer engine referral relationships, because there is this fear that you end up in a zero-click-world where nobody ends up clicking through to the full text and everyone just gets delivered answers in their AI tool. So you start to see usage really degrade. But there is also the possibility that this generates a new value that could drive traffic from places that we’re not currently realizing that value…We need to learn about what this new discovery path looks like.”
Some authors, Metzler acknowledges, may have concerns about their content being used in this way. She notes that the industry could do a better job of educating authors in this regard. “People hear AI and think of big tech companies gobbling everything up without permission or payment, using it to train their [large language models] and then capturing all that value for themselves and not giving any of that back. To be clear, I also think that’s terrible. I don’t think that’s good for authors. We don’t want that to be the expectation or the future.”
The Causaly deal is not AI training, but there is a misperception that everything involving AI is about training. “The first reason for authors to want us to participate in licensing on their behalf is because it is a way for us to push back against the granting of broad copyright exceptions for training, which we don’t think is the right thing for the creative industries and for academic publishing,” Metzler says.
A second argument, more specific to RAG, is that this is a part of discovery and the future of how content will be discovered. Metzler says: “What I would say to authors is, ‘we’ve all worked so hard, spent our lives producing all of this incredible science, and the way people are consuming that science is changing. AI is now going to be a part of that. If we want our content to be used and to be useful, and if we want to realize the benefits that are being promised from this ‘AI future’ that you may or may not have asked for, then we do need to participate.”
Metzler says her team is talking to start-ups that are competing with some of the big tech players. “I don’t want to see a future where there’s literally only three tools out there. I think that researchers should have choice between a range of different tools.” Domain-specific companies like Causaly are thinking about how to serve as a trust layer.
“It’s about discovery, access and trust. There are going to be more and more examples of those layers of trust being built around both trusted content, but also the technology layer that adds that additional layer of trust. People want to use AI, but they don’t want to trust their clinical decision making to Claude or ChatGPT,” Metzler says. “They want to be using tools that they think are more likely to be trusted.”
Causaly closed a $60-million Series B round in July 2023 and currently has more than 120 employees, including scientists with experience deploying AI in pharma R&D. Causaly’s core customer base includes leading biotech and pharma companies. Causaly says a dozen of the top 20 global pharmaceutical companies use its platform to accelerate drug discovery, including Novo Nordisk, Novartis, Takeda, Ipsen, and J&J.
Additional deals are likely in the wake of the Sage announcement. Earlier this year, Wiley struck a similar deal with OpenEvidence to bring Wiley’s medical content into the OpenEvidence platform.
