AI agents for nonprofits: CRM updates, grant reporting and volunteer intake
For the past couple of years, most of us have used AI to write things, like emails, summaries and social posts. In 2026 I expect a lot of nonprofits to start using AI agents that carry out the work as well.
Imagine Canada’s report, The State of Artificial Intelligence Adoption in Canadian Nonprofits, found most nonprofits are still in the “accessible, outward-facing” stage of adoption. We mostly use AI to talk about our mission. Very few of us use it to help run the mission day to day.
Here’s what that difference looks like in practice. Asking ChatGPT to summarize a meeting transcript is generative AI. An AI agent can join the meeting, catch every mention of a donor, update their CRM record, create follow-up tasks for the team and draft personalized thank-you notes in the background.
That takes the admin drag off your best people. Here are three places I’d start.
Three agents worth building
A CRM assistant for fundraisers
No fundraiser joined a nonprofit because they loved data entry. Keeping your CRM accurate still matters a lot, though. It’s your source of truth, and it’s often what separates a successful campaign from a missed opportunity.
The friction is the manual work of logging calls, mapping relationships and updating fields. An AI agent can act as your CRM’s invisible assistant. It reviews your sent emails, calls and calendar invites, updates the CRM and tracks the next step in each donor’s moves management plan, without anyone typing a word.
The report found only 27% of nonprofits use AI for fundraising. Taking the data entry off your gift officers frees them up to do what only people can do, which is build real relationships.
A grant agent for program staff
Grant management can feel like a game of don’t-miss-the-deadline. For program staff, the reporting load is close to a second full-time job, with hours spent going through funder portals and spreadsheets to stay compliant.
A grant agent lives in your grant folders. It monitors funder emails and portals, checks them against your current program data and drafts your monthly impact reports before anyone asks. The compliance checks happen every time, and your program leads get hours of their week back.
An intake agent for volunteers and participants
Can you give every volunteer or program participant a concierge experience? For most nonprofits, the honest answer is “we don’t have the staff.”
The friction here is the back-and-forth of intake, vetting and scheduling. An intake agent collects the information and then acts on it:
- It reviews a volunteer application.
- It checks the person’s availability against your needs.
- It sends a manager a green-flag alert when it spots a skill set you’re looking for, like legal or tech experience.
Only 8% of nonprofits use AI for volunteer recruitment and management, according to the report. Agents let you treat every supporter with the care you’d usually save for a major donor, and you can grow your impact without growing your payroll.
How to set up your first agent
- Pick one friction point. Choose the task your team complains about most, whether that’s CRM updates, grant tracking or intake.
- Write down every step a person does now. Include the small ones, like where they look, what they copy and who they tell. That list becomes the agent’s job description.
- Start with read and draft only. Let the agent read the emails, folders or applications and draft its updates. Have a person approve before anything changes in the CRM or goes to a funder.
- Decide what gets flagged. Pick the things that should go straight to a person, like a deadline coming up or a volunteer with a skill you need.
- Give it more access as trust builds. Once the drafts are reliably right, let it make low-risk updates on its own and keep approvals on the rest.
For leaders: plan for staff who supervise agents
The report shows a sector that’s experimenting, but still working in linear silos where one team waits on the next. The harder work this year is reorganizing around what agents make possible.
In a parallel organization, every staff member manages a small squad of AI assistants. You supervise the agent that does the data entry, and you approve the report it drafted from current data. Job descriptions shift from doing tasks to managing agents. I think the organizations that reorganize around this will be the ones that do best over the next decade.
A few questions for your leadership team:
- Which of our processes still wait on manual hand-offs?
- Where would an agent save our fundraisers or program leads the most time?
- Who approves agent work before it reaches donors, funders or the CRM?
- What training will staff need to supervise agents well?
If you want to find your team’s biggest friction points first, my free Nonprofit AI Readiness Roadmap is a good place to start.