Cambridge University Library is one of the oldest university libraries in the world. Today, it provides a home to more than ten million physical books, maps, manuscripts, photographs and other priceless objects – including the archives of Sir Isaac Newton, Charles Darwin and Stephen Hawking – across more than 35 locations.

Yet even the most historic of institutions are looking for ways to adapt to the rapid pace of digital change. That’s why the University’s library services launched its “Working Smarter” initiative in October 2025 to improve efficiency and innovation across the Librarys using AI.

The Library’s digital services team received funding through ai@cam’s AI for Ops programme, which supports AI solutions for operational challenges across the University. The initiative focuses on two areas:

• Using AI tools to unlock the University’s vast treasure trove of “hidden collections”

• Developing a chatbot to guide users through the Library’s archives.

“While we have a number of established workflows that have served us well,” explains Deputy Director of Digital Initiatives Lesley Gray, “we’re aware that ways of working are changing. We’re looking to see if we can use AI to augment our processes and automate some tasks, so we can work on areas where human expertise adds greatest value. This is not about replacing people – it’s about working smarter.”

Shining a light on hidden collections

One of the Library’s most pressing challenges is converting information from collections dating back hundreds of years into digital formats accessible to researchers worldwide. Currently much of this invaluable information is tucked away in rows of neatly written card catalogues, accessible only to scholars who have the chance to visit the Library in person. Before AI tools came along, transcribing these records would have taken countless hours of manual labour.

Building on progress made by the AI for Cultural Heritage Hub (ArCH) – which received seed funding from ai@cam – this latest digital initiative is exploring ways of converting handwritten text into machine-readable (MARC) catalogue data. The process involves imaging the physical cards, using AI to extract text, then converting the output into a bibliographic format that can be integrated into the library’s existing online catalogue.

The team will compare this proof-of-concept with other AI-enhanced approaches, such as an AI metadata assistant, to identify the most efficient way to illuminate these hidden collections.

Creating a Library chatbot

Another priority is developing an interactive, voice-enabled chatbot to help students and researchers navigate the University’s vast electronic collections. Drawing on existing information from the Library’s website, the chatbot will bring disparate details together to improve signposting to online resources. By automating responses to routine enquiries, the chatbot will free up staff time for more complex, strategic tasks.

Colleagues across the Library will be involved in developing the skills and frameworks needed to make sure that AI tools are deployed securely, ethically and sustainably – blending cutting-edge innovation with careful stewardship of knowledge.

Lesley explained that the ai@cam funding has allowed the team to set up a secure space to explore AI tools safely and responsibly – including for day-to-day tasks like recording minutes or analysing data.

“There’s a learning curve for everyone,” says Lesley. “We’re hoping that a lot more of our staff will improve their skills and understanding of how to get the best from AI.”

Sharing lessons learned

As the AI for Ops project enters its next phase, Lesley and her colleagues will focus on four specific case studies, with planning already underway for the first two. By March 2026, they expect to produce a series of reports and recommendations about how these AI tools can be rolled out.

Lessons learned will be shared widely – both across the University and externally with other libraries, through networks such as the Digital Scholarship Network (Research Libraries UK). In fact, there has already been significant external interest in the ArCH initiative, which has been showcased at the British Library and other national forums.

By blending innovation with careful stewardship, Cambridge University Library is ensuring that its vast resources remain accessible for generations to come.

Find out more about ai@cam’s AI for Ops programme and read more about the AI for Cultural Heritage Hub.