Catalyze AI · 2026 impact · West Michigan

Safe enough to try.

In 2026, 30 nonprofit leaders from 15 West Michigan organizations learned AI together in two cohorts, with their funder, Amway Corporate Philanthropy, at the table as a fellow learner. Here is what changed, measured against what the national research says the sector still needs.

Live on Zoom · Monday, November 16 · 2:00 to 3:30 pm ET Safe Enough to Try: Lessons from Frontier Leaders Implementing AI Hear these leaders tell it themselves. Free. Everyone who registers gets the follow-up resources.
Register free
30nonprofit leaders
15organizations
15/15organizations stayed to the end
7/7Fall organizations now have an AI policy written or in draft
North Kent ConnectLiteracy Center of West MichiganSTEM GreenhouseBridge Street MinistriesH.O.P.E. GardensFamily Promise of West MichiganAYA Youth CollectiveStorehouseCommunity Food ClubStreamsKids' Food BasketFeeding America West MichiganHome Repair ServicesWMCAT
Before and after

From unsure to ready

Both cohorts answered surveys at the start and the end. Pick a cohort to see how leaders moved.

Fall 2026: the same 8 leaders, start to finish

How to read this: each row shows the share of the 8 leaders who said yes, at kickoff (gray) and at the close (green). 100% means every one of them.

At kickoffAt the close

Leaders who answered both the kickoff survey (13 leaders) and the closing survey (10 leaders), matched by name: 8 people. Agree means "agree" or "strongly agree." No one moved backward on confidence. These are descriptive results from a small group, not proof the program caused every change.

Organizations7 of 7have an AI policy written or in draft. At kickoff, 1 had a written policy.
Organizations3 → 6of 7 have at least one leader using AI every day.
Leaders10 of 10at the close save at least an hour a week; 2 save more than 5.
Leaders9 of 10were very satisfied with the program.
Measured against the research

What the field is still asking for, these leaders did

Six gaps the national research names, from support to funding, set beside what happened in West Michigan.

01 Support · The research says
48% vs 30%

of employees are engaged when their manager actively supports their team's AI use, compared with those who lack that support. Gallup found manager support had the strongest link to engagement.

48%
30%
With manager support · without it
Gallup, 2026 · Reich and colleagues, 2026: psychological safety is the key to AI transformation
What happened here
15 of 15

organizations stayed from kickoff to Showcase, in both cohorts.

  • Every Fall organization sent a two-person team: one leader to own the why and one to own the how.
  • 11 of 14 Spring leaders at kickoff make or strongly influence their organization's AI decisions.
  • 8 of 10 Fall leaders took AI to their board or leadership team during the program.
  • Amway staff joined the sessions as learners, not observers.
02 Current use · The research says
92%

of nonprofits use AI, but 65% describe that use as reactive and individual. Among nonprofit staff, 30% still do not use AI at all.

92%
65%
Use AI · use it reactively and alone
Virtuous and Fundraising.AI, 2026 (346 nonprofits) · AI for Nonprofits Sprint (13,593 staff surveys, 387 organizations)
What happened here
31% → 70%

of Fall leaders using AI every day, kickoff to close (4 of 13, then 7 of 10).

31%
70%
Daily use at kickoff · at the close
  • Among the 8 leaders who answered both surveys, daily users went from 2 to 5.
  • Leaders at 5 of 7 Fall organizations estimate staff AI use rose 26 to 50% during the program.
  • At the Spring kickoff, 10 of 14 leaders had never used Copilot at work and 0 of 14 had used an automation platform.
03 Repeatable work · The research says
81%

of nonprofits use AI ad hoc, without documenting what works. Only 7% describe their use as strategic, with measurable mission impact.

81%
7%
Ad hoc, undocumented · strategic
Virtuous and Fundraising.AI, 2026
What happened here
12 of 12

organizations profiled below showed something reusable they built: a template, an agent, a policy or a process.

  • Site managers' notes became a 29-page activity manual in under an hour, now a template for other programs.
  • An email triage agent that quarantines instead of deleting and logs every decision.
  • One leader's Drive cleanup process became the cohort-wide Drive Cleanup prompt.
  • 10 of 10 Fall leaders at the close save at least an hour a week; 2 save more than 5.
04 Policy · The research says
76%

of nonprofit staff say their organization has no internal AI guidelines. Nationally, 47% of nonprofits have no AI policy at all.

76%
47%
Staff with no guidelines · organizations with no policy
AI for Nonprofits Sprint staff surveys, 2024 to 2026 · Virtuous and Fundraising.AI, 2026
What happened here
7 of 7

Fall organizations now have an AI policy written (2) or in draft (5). At kickoff, 1 had a written policy.

1/7
7/7
Fall organizations with a policy · kickoff to close
  • Fall leaders reporting no policy went from 8 of 13 to 0 of 10, the only change large enough to reach statistical significance.
  • 8 of 10 Fall leaders started drafting AI guidance themselves.
  • Policies were built with staff: one organization surveyed all 25 staff before writing a word.
  • Across both cohorts, at least 9 of 15 organizations have a policy written or in draft.
05 Skills and confidence · The research says
91%

of the least comfortable staff felt more comfortable using AI safely and ethically after a single hands-on AI 101 session. Staff learn best hands-on and from colleagues.

91%
Least-comfortable staff more comfortable after training
AI for Nonprofits Sprint, Fund for the City of New York · NTEN and The Bridgespan Group, 2026
What happened here
8 of 8

Fall leaders who answered both surveys now feel confident using AI and able to identify ethical risks, up from 5 and 4 of the 8 (63% and 50%).

4/8
8/8
Can identify ethical risks · kickoff to close
  • Leaders "unsure where to start" went from 4 to 0.
  • In Spring, leaders' average confidence using AI rose from 36% to 75% of a perfect score (3.6 to 7.5 out of 10), and it rose for every leader who answered both surveys.
  • Spring confidence leading AI capacity-building rose from 53% to 75%.
06 Funders · The research says
90%

of foundations offer grantees no support for adopting AI, even as more than 90% of organizations want to use it more.

90%
Foundations offering no AI adoption support
Center for Effective Philanthropy, 2025
What happened here
30 leaders

from 15 organizations, in two cohorts, funded by Amway Corporate Philanthropy, which invested in learning, not licenses, and sat at the table as a learner.

  • 9 of 10 Fall leaders were very satisfied; all 10 agreed the program fit the realities of nonprofit work.
  • 7 of 8 Spring leaders would recommend the program.
  • Leaders in both cohorts asked for a next phase together.
Who was in the room

15 organizations, in their own words

Highlights each organization shared at its Showcase. Time savings are self-reported.

Fall 2026

North Kent Connect

Surveyed all 25 staff before writing a policy built on a simple decision: Pause, then Proceed or Protect. They then used Claude to build volunteer onboarding, from a welcome video to a shorter path to a first visit.

Fall 2026

Literacy Center of West Michigan

Turned a year of talking into a quarterly priority with weekly check-ins, built a morning email digest, and kept AI inside Microsoft to protect learners' information.

Fall 2026

STEM Greenhouse

Wrote a first AI policy with the Policy Builder, took it to an all-staff meeting, and is standardizing hundreds of lessons so teachers save time.

Fall 2026

Bridge Street Ministries

Turned site managers' notes into a 29-page activity manual in under an hour, now a template for other programs, and wrote a formal AI policy.

Fall 2026

H.O.P.E. Gardens

A ten-minute AI-assisted draft became a funded $20,000 proposal, and a staff conversation led to a consent-first practice for photos of children.

Fall 2026

Family Promise of West Michigan

Built an email triage agent that quarantines instead of deleting and logs every decision, with client data kept out of every tool.

Fall 2026

AYA Youth Collective

Organized Drive and email with AI skills and connectors, starts each day with an AI morning brief, and built a self-paced case-management training for staff.

Spring 2026

Storehouse

Built its new website in-house with AI and prototyped a community resource-matching platform to show funders.

Spring 2026

Community Food Club

Turned recorded walkthroughs into visual procedures and a 34-page audit into a 15-minute podcast.

Spring 2026

Streams GR

Built donor research prompts and mapped its processes, and reports 10 to 15 hours a week saved on donor work.

Spring 2026

Kids' Food Basket

Gave its leadership team AI tools with a 30-day challenge, then moved to an acceptable-use policy, data governance and staff training.

Spring 2026

Feeding America West Michigan

Wrote an organization-wide AI policy framework for its board and training protocols for its fundraising team.

Also in the Spring 2026 cohort: Home Repair Services WMCAT and a national youth and community development organization planning listening sessions with 90 of its directors.

What we learned

Trust comes first, then the tools

01

Ask before you write the rules

Policies built from staff voice earned trust. Staff asked for clear guidelines, practical training and privacy safeguards.

02

Privacy is a design choice

Some information never goes into a tool. Leaders decided that on purpose: client data out, consent first, keep it inside the walls.

03

The repeatable thing is the win

A saved prompt, a template or a logged agent keeps paying off long after the session ends.

04

Structure beats willpower

A date, a peer who will ask, and a room to show your work did more than good intentions.

05

The funder in the room changes the room

With Amway learning alongside them, it became safe to try, to fail and to learn.

06

Learning together works

Half of the Spring leaders asked, in their own words, for more in-person, hands-on time together.

We move at the speed of relationships, and you can't build efficiencies into relationships. It's helped us identify what's uniquely human in our work.

Executive director, Fall 2026 cohort

This cohort provided a support and accountability structure for me to engage with this technology when I otherwise probably would not have.

Senior director of programs, Fall 2026 cohort
Monday, November 16 · 2:00 to 3:30 pm ET · Zoom

Hear it from the leaders themselves

Safe Enough to Try: Lessons from Frontier Leaders Implementing AI. Part of the Women Build Safety Challenge 2026. Free. Everyone who registers gets the follow-up resources.

Register on Luma
Register for Nov 16