Hi Linda,

This is a question we think about constantly. Quick disclosure: I co-founded Atom Grants, which works with 50+ research institutions on exactly this, so I have a point of view but I'll try to keep it useful rather than promotional.

A few patterns from what we're seeing work across partners like NYU Langone, Auburn, and Denver Health:

Funding identification is where AI delivers the clearest ROI today, but only if it's automatically pushing to faculty, doing semantic matching against a researcher's actual interests rather than manual keyword search. The bar has moved.

Proposal development is the area with the most hype and the most disappointment. What actually works is using AI to generate structured guides from solicitations and to red-team drafts against review criteria, not to write the proposal itself. Faculty ownership of the narrative matters.

The biggest mistake we see offices make is evaluating AI tools in isolation rather than asking how they fit into the faculty workflow end to end. A brilliant tool that faculty don't adopt is worse than a mediocre one they use weekly.

Happy to chat off-list if helpful, and I can share a SRAI Catalyst piece we co-authored with Denver Health on their implementation experience if anyone's interested.

Best,

Tomer du Sautoy
Co-Founder & CEO
Atom Grants