More than 100 Agentforce use cases for charities, spanning fundraising, programmes, marketing, volunteering, grantmaking and operations. Find your first agent.
This week we launched our new website, and with it a resource we have wanted to share publicly for some time: the Agentforce use case library, a catalogue of more than 100 ways artificial intelligence agents can work for charities and nonprofits.
This post explains what is in the library, why we built it, and how to use it to find the right first agent for your organisation.
First, what is an agent?
Agentforce is Salesforce’s platform for building artificial intelligence agents. An agent is a piece of software that completes a defined job using your Salesforce data: summarising a donor’s history before a meeting, drafting a stewardship report, or answering a supporter’s question about their regular gift. It works within the permissions you give it, and the best implementations keep a person reviewing anything that matters.
That definition is deliberately unglamorous. An agent is not a replacement for your team. It is a way of handing over the repetitive, time-consuming parts of a role so the humans can spend more time on the parts that need a human.
Why we built a library
Almost every conversation we have about artificial intelligence starts with the same question: “what could an agent do for us?” The honest answer is “a great many things”, which is true and completely unhelpful. Teams do not need to be told the technology is capable. They need a shortlist of jobs, relevant to their function, that they can weigh up against their own workload.
The wider sector data backs this up. As we covered in our look at the Charity Digital Skills Report 2026, 79% of charities now use artificial intelligence routinely, yet most of that use is informal: drafting emails, tidying meeting notes. The missing step between experimenting and strategy is knowing which specific jobs to give the technology. That is what a use case library is for.
So we catalogued the use cases we scope with charities, in one place, in plain language. Every Discovery and Planning engagement we run now includes a use case mapping session against this library.
79%
of charities now use artificial intelligence routinely (Charity Digital Skills Report 2026)
How the library is organised
The library is filterable by the six functions where charities spend most of their time:
Fundraising, from prospect research through to stewardship
Programme management, supporting caseworkers and service delivery teams
Marketing, for supporter communications and journeys
Volunteering, from recruitment through to recognition
Grantmaking, for organisations that distribute funds as well as raise them
Operations, the internal work that keeps everything else running
Each use case has a short description of the job the agent does and a practical example of how a nonprofit would apply it.
A few examples to give you a flavour
Prospect Research Agent (fundraising). Pulls together what is already known about a prospective donor, from your own records and public sources, into a briefing your fundraiser would otherwise spend hours assembling.
Post-Event Follow-up Agent (fundraising). Drafts personalised thank you and follow-up messages for event attendees while the event is still fresh, ready for a fundraiser to review and send.
Stewardship Reporting Agent (fundraising). Assembles the update a major donor or funder expects, drawing on programme data you already hold, so reporting deadlines stop eating whole weeks.
Participant Management Agent (programme management). Summarises a participant’s notes, goals, support history and upcoming tasks, so a caseworker picking up a file can get up to speed in minutes rather than an afternoon.
Audience Segmentation Agent (marketing). Builds and refines supporter segments for campaigns using the data you already hold, instead of a marketer exporting spreadsheets and hand-building lists.
Donor Support Agent (fundraising). Answers routine supporter questions, such as updating a regular gift or requesting a receipt, and hands anything sensitive straight to a person.
There are more than 90 others, and browsing by your own function is the best way in.
How to choose a first use case
More than 100 options can be its own problem, so here is the test we apply in use case mapping sessions. A strong first use case passes all four:
The job is defined and repeatable. “Summarise this participant’s file” is a job. “Help us with fundraising” is not.
The data the agent needs already lives in Salesforce, and is in reasonable shape. An agent working from incomplete or duplicated records will produce confident nonsense.
A person stays in the loop. Early use cases should draft, summarise and prepare, with a human reviewing before anything reaches a supporter or participant.
You can measure the result. Hours saved per week, response times improved, income influenced. If you cannot measure it, you cannot make the case to do more.
Use cases that pass the test tend to be unglamorous and immediately valuable. That is a good sign, not a bad one.
What the library will not tell you
A library of possibilities is not an implementation plan. It will not tell you whether your data is ready, what the right licensing looks like for your size of organisation, or how to set permissions so an agent sees only what it should. After the Beacon incident reminded the sector how much supporter data deserves protecting, that last point matters more than ever: an agent inherits the access you give it, so scoping access is part of scoping the agent.
Those questions are what our Discovery and Planning engagements answer. We map your shortlist against your data, your team and your budget, and we will tell you plainly if a use case you like is not ready to build yet.
Start browsing
The full library is open to everyone, no form required. If you find two or three use cases that would change your team’s week, book a discovery call and we will help you work out which one to build first.