Benchmarks & insights.
How purpose-built AI compares to enterprise tools and manual research. Open benchmarks, technical deep-dives, and lessons from the field.
Benchmarks
PIF-Bench v2, our open, rubric-graded benchmark — then Rōmy against enterprise wealth screening and manual prospect research.
PIF-Bench v2 — Live track
Donor reports each product wrote on April 25, 2026 for four real philanthropists, graded on 59 dated, sourced checks. No product met every check for any prospect.
PIF-Bench v2 — Packet track
20 tasks across prospect research, grants, gift compliance, major gifts, donor data and finance (259 checks), one attempt per model.
Time per Prospect
From first search to actionable report
Cost per Report
Fully loaded cost per full research report, depending on subscription tier
Annual Cost
For a team researching ~200 prospects / month
Report Depth
Sections in a full donor intelligence report
Blog
Technical deep-dives, research findings, and perspectives on nonprofit fundraising.
Want to run your own benchmarks?
Try Rōmy on your own prospect list. See how purpose-built donor intelligence compares to your current workflow.