Not For Profit Shared Interest Group

Recent industry data shows over 90% of non-profits are experimenting with AI tools (for grant writing, data analysis, or donor engagement), yet fewer than 10% report seeing major, needle-moving improvements. Many finance teams are stuck between curiosity and implementation risks.<sources-carousel-inline ng-version="0.0.0-PLACEHOLDER" _nghost-ng-c2510184733=""> </sources-carousel-inline>

Where is your organization on the AI adoption maturity curve? Beyond basic administrative tasks, what guardrails, data governance policies, or internal controls is your finance team putting in place to ensure AI delivers real operational value without creating security or compliance risks?

------------------------------
Rinku Bhattacharya, DPS, CMA, CPA
CFOO, Boys and Girls Club of New Rochelle
Rinku_Bhattacharya@...
Hartsdale, NY 10530
------------------------------

Our organization just addressed AI officially in June of this year. Even now we have recommendations on use, but no official policy. I've personally been using AI to help with financials and data analysis since spring of 2025. 
I've expressly told my team to use only the company-provided AI resources of CoPilot, as part of our MS Office suite, and our enterprise version of Claude. They are absolutely not to use a personal AI and not to use a free version of AI. 

------------------------------
Patricia Howard CMA
Controller
Pleasant Pln OH
------------------------------

This is the right conversation to be having, and Patricia's point about restricting to enterprise-grade, company-provided tools is the essential first control - especially in nonprofits handling donor and grant data, where a free tool can mean data leaving your environment entirely.
I'd add a second layer that's easy to miss. Tool policy governs the input - who can use what. But the harder question is the output: once AI touches a financial process, who signs off on the result, and can they explain how it was produced? In grant-funded and multi-donor environments, that's not a technology question, it's an audit-defensibility one. An AI-assisted allocation that no one can reconstruct six months later is a finding waiting to happen, regardless of which tool produced it.
The organizations I see handling this well pair a tool policy (Patricia's point) with a simple accountability rule: no AI-influenced number enters the financials without a named human who owns it and can defend the reasoning. The first protects your data. The second protects your audit.
Rinku, to your maturity-curve question - most nonprofit finance teams I work with are still at the tool-policy stage. The accountability layer is where the next year of maturity happens.

------------------------------
Marina Ter Sargsyan
Chief Financial Officer
Germantown MD
United States
------------------------------