AI for Non Profits Network: Weekly Briefing 03/10
The weekly digest from a network of non-profits 92%: of US nonprofits are using AI only 7% have changed what they can do. This week: what separates them and what the governance gap means for your team
In The Briefing this week:
đ Whatâs Caught our Eye: The 7% Problem
đ Thought for the Week: The Governance Gap and Learnings from Our Workshop Last Week
đ Interesting News & Funding Calls
đ From Across the Network
1) đ What Caught Our Eye: The 7% Problem
A new benchmark study released this week captures something many of us have been watching: AI use has become nearly universal across the nonprofit sector, but the gains remain surprisingly thin.
The 2026 Nonprofit AI Adoption Report, drawn from 346 US nonprofits, finds that 92% are already using AI-enabled tools in some form. The number that keeps staying with me: only 7% have meaningfully improved their ability to achieve their mission.
The report calls this the efficiency plateau. Organizations are drafting content more quickly, routing queries more efficiently, automating tasks that used to take hours. But the work AI is touching rarely reaches strategy, budgeting, prospect research, or anything that will outlast the individual using the tool. It accelerates what was already happening. It does not expand what is possible.
Speed gains are visible and immediate. A staff member drafts a grant summary in ten minutes instead of ninety. The time saving is real. What does not appear in that calculation is whether the organization is now applying for better-fit grants, reaching more donors, or evaluating programs more rigorously. Those questions are slower and harder to attribute. So âwe are using AIâ becomes the answer to a question nobody is quite asking.
81% of respondents in the report are using AI tools without documenting their workflows. The knowledge of what works - which prompts, which outputs feed into which decisions, what the quality checks look like - lives inside individual staff members. When those people leave or move roles, the capability goes with them. It never belonged to the organization. And 47% of nonprofits in the sample have no AI governance policy at all. Donor records, beneficiary data, and confidential program information are passing through consumer AI tools with no formal framework for what that exposure means or who is accountable when something goes wrong. We hosted a network-wide workshop on exactly this last week - If you want the tools drop us a message at hello@aifornonprofitsnetwork.org.
The 7% who have crossed into mission-level transformation share a common pattern. They have moved from individual AI use to organizational AI design. They have documented what they do with AI, assigned accountability for how it is used, built systems that persist beyond any one person, and created the conditions to evaluate whether any of it is actually working. The capability belongs to the organization, not to whoever happens to know the right prompt today.
This connects to a pattern we flagged in our February 10 briefing: 85% of nonprofits were experimenting with AI tools but only 24% had established a formal roadmap. The efficiency plateau is what that gap looks like six weeks later, measured at scale. The full report is worth reading carefully. It is one of the more rigorous datasets the sector has produced on AI adoption to date, and the efficiency plateau it describes is something we will keep returning to in this briefing.
Our Next Workshop: Leading on AI When You Feel Behind - Join Us 1st April
If youâve spent the last six months nodding along in AI conversations while quietly wondering whether everyone else understands this better than you do this is the session is for you.
Our free 90-minute working session is designed specifically for nonprofit leaders who need to make smart decisions about AI without understanding every detail of how it works. Feeling âbehindâ isnât a sign of limited capability. Itâs a rational response to an irrational pace of change.
Joining us is Paul Butcher, from CommonSensing AI and former CMO and Digital Lead at Save the Children. Youâll walk away with a simple framework for AI decisions, a draft governance position for your board, and clarity on where to start - and what to stop worrying about.
Wednesday April 1 | 2pm EST | 90 minutes | Free
Reserve your spot by replying to this email or writing to us at hello@aifornonprofitsnetwork.org.
2) đ Thought for the Week: The Governance Gap and Learnings from Our Last Workshop
What does it take for AI use to become organizational capability rather than individual habit? The reportâs answer - organizational readiness - is accurate but underspecified. Here is what I think it actually means in practice.
The Capability Ladder connection
For readers who have been following this briefing for a while, the AI Capability Ladder framework we introduced in November 2025 maps directly onto what the report is measuring. The Ladder describes a five-rung progression: from ad hoc individual AI use at Rung 1, through documented and governed use at Rung 2, to integrated team workflows at Rung 3, system-level AI design at Rung 4, and fully mission-integrated AI at Rung 5.
The efficiency plateau is what Rung 1 looks like when it reaches 92% penetration. Almost every organization has someone using AI for something. Almost none of them have climbed the first rung. The 7% who have reached mission-level impact are at Rung 3 or above. The gap between them and the majority is not capability. It is structure.
Workflow documentation
The first structural move is documentation, and it is simpler than most organizations assume. For each recurring AI-assisted task in your organization, someone writes down five things: what the task is, what tool is used, what the input looks like, what the output feeds into, and who reviews it before it is acted on. That is a shared document. It takes an hour to produce for most tasks. A morning to cover an organizationâs full AI footprint.
The result is institutional knowledge that survives staff turnover. An organization that has documented its AI workflows can evaluate whether they are working, train new staff to replicate them, and build on them deliberately over time. An organization that has not documented them has a collection of individual habits that will evaporate the moment those individuals move on. The 81% figure is not a technology problem. It is a knowledge management problem with a straightforward solution.
Governance policy
The second structural move is a governance policy, and again the bar is lower than most nonprofit leaders assume. A basic AI governance policy for a small to mid-size nonprofit answers three questions: what data can be fed into AI tools, who has authority to approve new AI use cases, and what happens if something goes wrong. We developed something in a Governance workshop last week - email us if you want the policy.
The 47% without any governance policy are not primarily at risk from dramatic AI failures. They are at risk from the ordinary cumulative exposure of staff running donor data through consumer tools, making decisions based on AI outputs without review, and building institutional dependencies on services they have no contractual relationship with. That is a slow risk that compounds quietly. A policy changes the exposure profile significantly. The concept of Shadow AI - staff running donor data through consumer tools before any policy exists - came up throughout our February 4 workshop briefing. The 47% without governance is not an abstract risk. It is what Shadow AI looks like when measured at scale across 346 organizations.
The structural move
Put these two things together and what you have is the foundation for climbing the Capability Ladder. Documented workflows mean you can evaluate what you are doing. A governance policy means you have defined accountability for how it is done. Those two things create the conditions for genuine improvement: you can see what is working, you can change what is not, and the knowledge belongs to the organization rather than to whoever happens to be in the building today.
For teams at the efficiency plateau right now the next move is not a new tool. It is a document and a meeting. A morning to map your AI workflows. An afternoon to draft the governance principles. That is the on-ramp to the 7%.
What is the single AI-assisted task your team relies on most heavily, and could you describe exactly how it works to a new staff member starting tomorrow?
3) đ Interesting News
Nonprofits Embrace AI, But Little To Show For It So Far (The NonProfit Times)
How AI Will Transform Nonprofit Operations in 2026 (Daxko)
AWS Imagine for Nonprofits (AWS)
4) đ From Across the Network
The Nonprofit Fundraisers Symposium runs March 25-27 in Washington, D.C. Three-day program covering fundraising strategy, AI in direct response, and practitioner peer learning. Organized by the Direct Marketing Association of Washington and The Nonprofit Alliance Foundation.
Have an event, case study, gathering or interesting insight you would like to share with the network? Drop us a note by replying to this email.
Weâre a network of non-profit leaders interested in how AI is impacting our sector and driving mission. Youâre receiving this email because you expressed an interest in the network. If you think this was a mistake, you can unsubscribe using the link below.