Digital Science opens £25,000 grant for trusted AI workflows

Digital Science has opened applications for its 2026 Catalyst Grant, which will award up to GBP £25,000 for early-stage projects focused on trusted agentic AI workflows.

The grant is open worldwide to individuals, start-ups and research teams, with equity-free funding available. Applicants do not need revenue, a finished product or a full business plan.

This year's theme is "Agentic Workflows You Can Trust", aimed at software ideas that use AI agents in research settings with clear audit trails, governance and provenance. The programme is seeking projects that address any part of the research lifecycle, including authorship and writing, institutional research processes, funding and decision tools, and publication workflows.

Now in its 16th year, the initiative is part of Digital Science's broader effort to support early-stage technology ideas in research. Applicants may submit a prototype, a working product or a developed concept, as long as they explain the problem, their approach and why the work fits the theme.

Focus on trust

Steve Scott, Vice President, Portfolio Development, at Digital Science, said the company is seeing a shift from earlier generative AI tools to systems that can carry out multi-step tasks.

"AI tools are increasingly able to act on research, not just describe it. The next breakthroughs will come from agentic workflows, autonomous and multi-step, that researchers, institutions and funders can actually rely on. We're funding the early-stage builders making AI agents safe to put in front of an institution: autonomous in service of researchers, with provenance, governance and accountability built in.

"This is a different category from the generative AI tools we were funding three years ago. An agentic workflow plans, executes and reviews multi-step work rather than answering a single prompt. Trust in this case means quality rather than detection. An agent that shows its working, not a tool that catches bad actors after the fact," Scott said.

The emphasis reflects a broader debate in research and academic publishing over how AI systems should be introduced into sensitive workflows. Universities, funders and publishers are weighing how to use automation while preserving oversight, record-keeping and confidence in outputs.

By framing the grant around trust, Digital Science is emphasising systems that can show how decisions or actions were reached. In practice, that points to projects with transparent processes and clear controls, rather than tools focused only on generating answers.

Early-stage support

The award is worth up to GBP £25,000 for the most promising ideas. The scheme is intended for early-stage work, including concepts still being shaped as well as products already in development.

That lower barrier to entry may broaden the applicant pool beyond more established start-ups. Eligible applicants include university research teams and independent developers, alongside young companies building AI tools for scholarly and institutional use.

The programme is designed to support novel uses of technology that aid research and its wider social impact. Digital Science's portfolio includes products and services used across the research sector, from publishing and manuscript preparation to analytics and institutional systems.

Beyond funding

Scott said the programme's value extends beyond the cash award.

"Catalyst Grant's real value lies beyond the funding itself. It's in the opportunities provided to winning teams: the conversations it starts, practical advice, introductions to other experts in the field, and a sharper idea of what successful innovation looks like," he said.

The grant adds to a growing number of schemes directing small pools of funding to AI-related research tools, particularly where larger investors may view products as too early or too specialised. In Digital Science's case, the focus remains closely tied to the needs of researchers, institutions and funders rather than general consumer AI applications.

The application window closes in early October, with entries invited from around the world for projects that can show how agentic AI can be applied to research work with accountability built in.