Agent Build Guide for Nonprofits · Updated
Mastering AI in Your Nonprofit
Automate the routine, keep the relationship. You do not hand your whole development shop to an agent. You pick one repeatable task, build one agent for it, and keep a human in the loop so the final connection with the donor stays with you. This guide covers what to automate first, what a philanthropy agent is made of, when to build versus when to just write it down, five paste-ready builds, and what to guard against.
Start the two-minute assessment01 · What to automate first
What should I automate first?
Automate the routine you already do the same way every week. Rule-based, repeatable, low-ambiguity tasks make the best first builds, because you can tell inside a single run whether the agent got it right or not. The cultivation itself, the call, the coffee, the ask across the table, is never the candidate. The work around it is.
Three questions, about two minutes. Every answer routes to a specific build.
Question 1 of 3
Question 2 of 3
Question 2 of 3
Question 3 of 3
Your starting build
Answer the three questions and the one agent to build first appears here, with a direct link to its recipe. Answer all three for the sharpest match.
If you would rather just read the routing:
| What eats your week | Start here | Why |
|---|---|---|
| Drafting the same kinds of donor updates and appeals | Campaign Sprint | It drafts the donor email. You still read it and hit send. |
| Rewriting one impact story for every channel | Impact Storyteller | It does the reshaping. You keep the story. |
| Researching a donor or funder before a meeting | Prospect Brief | It hands you the one-page brief before you walk in. |
| Moving gifts, RSVPs, and contacts into the CRM by hand | CRM Admin | This is the job agents are best at. Start here. |
| Reading foundation guidelines to find a fit | Grant Prospector | It takes the first pass so you only read the foundations that actually fit. |
| None of these, mostly cultivation and conversation | Grant Prospector | It only reads and summarizes, so it is the lowest-risk agent to try. If gift entry is more your edge than reading, start with CRM Admin instead. |
Take the builds with you
Want these five builds in your inbox?
Drop your email and I will send you all five agent builds, paste-ready, with your starter one in full. Set the first agent up tonight and keep the rest for when you need them.
One email with the builds. Reply anytime and it reaches me directly.
02 · The parts
What is a philanthropy agent actually made of?
A Gumloop agent is an assembly of six kinds of parts, a Trigger, Connectors, Skills, Knowledge Sources, Subagents, and Abilities, rather than one large model that somehow already knows your mission. Translated into your language, the technology stops feeling like a black box.
Trigger
The heartbeat that starts the agent running: a schedule, an event in a connected app, or a check you describe in plain language. For a giving sprint it might be 8:00 AM every Monday to check campaign progress. Gumloop offers three kinds: Create With AI for custom polling Triggers you describe yourself, Event-Based Triggers built for a single service, and Scheduled Triggers that run on a set schedule or once at a set time.
Connector
An authorized link to an outside tool the agent reads from or writes to. Gumloop documents more than 150 integrations, including Gmail, Outlook, Google Calendar, Sheets, Docs and Drive, Slack, Notion, Airtable, HubSpot, Salesforce, LinkedIn, Instagram, YouTube, Canva, Gamma, Beehiiv, Stripe, QuickBooks, Cal.com, Luma, Zoom, Google Analytics, and Google Search Console.
Skill
A reusable set of instructions, and optionally templates and scripts, that teaches an agent how to do one specific task. You can build one three ways: Create With AI, Write Skill Instructions, or Upload Files.
Knowledge Source
Your organization's brain: impact reports, strategic plans, past successful grant applications, your case for support. Gumloop's Brain connects Notion, Google Drive, Slack, GitHub, Confluence, Zendesk, and Salesforce, plus uploaded files. Attach knowledge and the agent gains a Search Company Brain tool that returns relevant snippets with citations instead of guessing.
Subagent
A separate agent your agent calls to handle a distinct job with its own tools. An agent can clone itself, which appears in selection lists as (Me), with a depth limit of one so clones cannot spawn clones. It can also invoke a different named agent, with no depth restriction.
Ability
A built-in capability the agent has without connecting anything: Web Search, Web Fetch, Image Generation, a Code Sandbox that runs Python and shell commands, and Ask Question, which pauses and asks you for input through a structured choice card.
Walking the builder screen
Agent Preferences is where a build starts. Pick the model that powers the agent, then use the Add instructions for the agent box to write what it is, who it works for, and how it should behave. This box is the agent's job description, and most of what people expect from a Skill actually belongs here.
Beneath the instructions sits the Allow Self-Updates toggle, which lets the agent revise its own instructions. It is on by default. Leave it on and you can correct the agent in conversation, tell it to update its instructions so it always works that way, and watch the change land in the box you just wrote.
Triggers is where you decide what starts it. The panel carries an AI Managed setting, and the agent can create and manage its own Triggers during a conversation, so scheduling can be a thing you ask for rather than a form you fill in.
Connectors is the authorization step, shown with an AI Discovery setting. Add only what this agent needs. An agent holding every Connector in the company is the signal that you should have split it into specialists.
Skills holds the reusable procedures, with an AI Skill Editing setting on the panel. Skills can be updated by the agent itself when you correct it, or edited by hand by anyone with agent editing access.
Knowledge Sources is where you attach the documents and workspaces the agent should read before it answers. Subagents is where you add the other agents it can call, including its own clone as (Me). Abilities holds the built-in switches, Web Search and Web Fetch among them.
Now see how the parts combine03 · Build it or write it
When do I create a Skill, use a Subagent, or just write it into the instructions?
Write the logic into the instructions when it governs everything this one agent does and fits in a few lines. Make it a Skill, a reusable set of instructions that teaches an agent one specific task, when it is a standing procedure you would otherwise paste in as a long document, or when a second agent will need the same procedure. Hand it to a Subagent, a separate agent called to do a distinct job with its own tools, when the work is a different job with its own steps, its own connections, and its own back-and-forth.
Gumloop's documentation states the same test more bluntly:
"If you are about to paste a long SOP into the agent instructions, make a skill. If you are about to give one agent every connector in the company, split specialists and call them as subagents."
Gumloop documentation
| Object | Build it when | What it is | Where it runs |
|---|---|---|---|
| Instructions | The rule applies always, to this agent only, and fits in a few lines | The text in the Add instructions for the agent box | Every turn, automatically |
| Skill | The procedure is long, repeatable, or shared across agents | A reusable set of instructions, plus optional templates and scripts | Inside the current conversation, when relevant |
| Subagent | The work is a separate job with its own tools and its own reasoning | A different agent, or a clone of this one shown as (Me) | In its own conversation, then reports back |
| App Rules | An action must be impossible, not merely discouraged | An enforcement policy, set by agent editors or org admins | At the tool call, before it executes |
App Rules are the fourth object, and the one people skip. Skills suggest, App Rules enforce. Gumloop's own comparison page puts it this way:
"A skill can ask the agent not to delete Slack messages. An app rule can stop the delete call."
Gumloop comparison page
Worked example: Campaign Sprint, instructions
The Campaign Sprint agent drafts donor updates during a giving push. "Draft, never send. Match the warm, jargon-free voice in the Draft Donor Outreach Skill. Never use guilt or pressure. Stop and ask before touching a major-gift donor." That is a few lines, it applies to every draft the agent produces, and it belongs in the instructions box. Building a whole Skill for it would add a file to maintain and change nothing about the output.
Worked example: Impact Storyteller, a Skill
The Impact Storyteller goes the other way. Every format you publish carries real rules about length and tone, and how a 50-word social post differs from a formal grant update, and writing all of that into the instructions box means pasting a document into a field meant for a job description. That is a procedure, and you will want the same one the day you build a second agent for another channel. Make it a Skill.
Worked example: Prospect Brief, a Subagent
Prospect research does not belong on the Campaign Sprint agent. Research needs the web, your CRM, your calendar, and your notes, and bolting all of it on leaves you with one overloaded agent holding every Connector you own. Keep research as its own agent with its own connections. The Campaign Sprint agent calls it as a Subagent when a draft needs context, and it hands back the brief.
The CRM Admin agent is where App Rules earn their keep. It writes to your live donor database, so the rule that stops a delete or an overwrite has to be enforcement, not a politely worded instruction.
Jump straight to a build: Campaign Sprint, Impact Storyteller, Prospect Brief, CRM Admin, or the Grant Prospector. Not sure which fits your week? Take the assessment.
See the five buildsThe builds
Five agents to build first
These five are common, bounded, low-risk jobs, and each one sits in a different function of running a development shop: your appeals, your storytelling, your donor research, your database, and your grant pipeline. None of them lets the agent send to a donor, publish, submit an application, or spend anything on its own.
Each recipe uses the same ten fields so you can compare them before you pick one. The instructions block in each is written to paste straight into the "Add instructions for the agent" box in Agent Preferences, with your own details swapped in. Each recipe also names the Skills to build and why; the test is the one from the decision table: a reusable process becomes a Skill, a one-off detail stays in the instructions. New to the parts vocabulary? The anatomy section defines all six in a minute of reading. Build one, watch it run, then build the next from what you learned.
When you're ready, here are the other four builds.
01/5
04 · Campaign Sprint · today's live build
Campaign Sprint
This is the agent you build in today's session. It reads your live campaign tracker, drafts a warm, personalized update to each donor who is ready to hear from you, and waits for your yes before anything sends. It is for the fundraiser running a giving sprint who is doing the math and the drafting by hand, one donor at a time.
Drafts your donor outreach. It never sends one.
04 · Campaign Sprint · today's live build
Campaign Sprint
This is the agent you build in today's session. It reads your live campaign tracker, drafts a warm, personalized update to each donor who is ready to hear from you, and waits for your yes before anything sends. It is for the fundraiser running a giving sprint who is doing the math and the drafting by hand, one donor at a time.
Drafts your donor outreach. It never sends one.
Connectors, and what each one does here
- Google Sheets. Reads your campaign tracker: the donor list, each donor's last gift and date, the campaign goal, and the amount raised so far.
- Gmail or Outlook. Saves each donor draft where you will see it, as an unsent draft, and never sends. Gmail's documented capability includes managing drafts, which is the basis of this recipe.
Skills to create, and why
A Skill is a reusable instruction pack an agent follows for a specific task. Build one: a Draft Donor Outreach Skill through Write Skill Instructions. It holds your tone and voice rules, warm and conversational, plain language, no jargon, no guilt or pressure, and the exact shape of a good outreach email: greeting, gratitude for the last gift, the campaign update, a soft ask, a warm close. Give the room three minutes to make the tone their own. A youth arts group writes "energetic and vibrant." A medical research fund writes "empathetic and grounded."
Setup steps in order
- Create a new agent and name it for the campaign, not for the tool.
- In Agent Preferences, choose your model and paste the instructions block below into "Add instructions for the agent."
- Add the Google Sheets Connector, then Gmail or Outlook.
- Create the Draft Donor Outreach Skill and rewrite the tone lines to match your organization.
- Run the agent manually against three real donors from your tracker and read every draft.
- Set the Scheduled Monday 8am Trigger only after those three drafts pass.
Make it yours, then paste it in
Nothing you type here leaves your device. It only rewrites the block below.
Paste this into "Add instructions for the agent"
You draft donor outreach for our current campaign sprint. You never send. Every Monday at 8am, read the campaign tracker. For each donor marked ready in the tracker, draft a warm, personalized email that thanks them for their last gift, shares where the campaign stands against its goal, and softly invites them to give again. Follow the Draft Donor Outreach Skill for tone and structure. Keep each draft under 150 words. Do not use exclamation points. Do not use em dashes. Use only the facts in the tracker: the donor's name, their last gift and date, the campaign goal, and the amount raised so far. Do not invent an impact story, a number, or a past conversation. Never use guilt, urgency, or manufactured panic. We build relationships, we do not demand transactions. Save each draft as an unsent draft for me to review. Never send. Never delete or archive anything. If a donor is a major gift, a lapsed donor, or anyone flagged for a personal note, stop and use Ask Question to show me the record and ask how I want to handle it.
Open Gumloop and paste this block in. Building it live with us today? Bring it to the session. Stuck partway through? Email hello@bnedai.com.
You're done when
02/5
05 · Impact Storyteller
Impact Storyteller
This agent takes one finished impact story and reshapes it into every format you need, a newsletter blurb, a short social post, a formal grant update, without adding a claim you did not write. It is for the person who writes a program's story once and then rewrites it five times for five audiences.
Drafts every version. It never publishes one.
05 · Impact Storyteller
Impact Storyteller
This agent takes one finished impact story and reshapes it into every format you need, a newsletter blurb, a short social post, a formal grant update, without adding a claim you did not write. It is for the person who writes a program's story once and then rewrites it five times for five audiences.
Drafts every version. It never publishes one.
Connectors, and what each one does here
- Google Drive. Watches the folder where you drop a finished impact story.
- Google Docs. Reads the source and writes all output into one new document you review.
- Beehiiv. A verified Connector for newsletters, documented for publishing posts and managing subscribers. Sending to your list is exactly what you do not hand over on a first build, so add this Connector only when you are ready and keep the publish step behind your own yes.
Canva and Gamma are verified Connectors too, if a story needs to become a donor graphic or a board deck. Leave both out of your first build.
Skills to create, and why
Build a House Voice Skill through Write Skill Instructions, specifying the exact length and tone of each output: newsletter blurb length, social post length, the shape of a grant-update paragraph, and a plain-language rule that bans nonprofit jargon like "capacity building" or "leveraging synergies." Build a second Source Discipline Skill through Create With AI, stating that every claim in an output must already exist in the source story. This is the guardrail that keeps the AI from inventing a rescued puppy or a well that was never dug. Gumloop supports a faster path for both: do the task once in chat, correct it until it is right, then tell the agent to turn that into a Skill.
Setup steps in order
- Create the agent and paste the instructions block below into "Add instructions for the agent."
- Add the Google Drive and Google Docs Connectors.
- Create a folder for this purpose and put one finished impact story in it.
- Create the House Voice Skill with your real lengths and tone.
- Create the Source Discipline Skill.
- Run the agent manually on that story and trace every claim in every output back to the source.
- Set the Create With AI Trigger to watch the folder once the manual run is clean.
Make it yours, then paste it in
Nothing you type here leaves your device. It only rewrites the block below.
Paste this into "Add instructions for the agent"
You turn one finished impact story into channel-ready drafts. You never publish. When a new document appears in the Impact Stories folder, read the whole thing before you write anything. Produce all output in one new Google Doc titled with the source title plus "repurpose drafts." Include a 200-word newsletter blurb, a 50-word social post, and a formal grant-update paragraph. Follow the House Voice Skill for the tone and exact length of each. Use only the facts, numbers, names, and outcomes that already appear in the source story. Do not add a statistic, a beneficiary story, or a quotation that is not in the source. Strictly avoid nonprofit jargon like "capacity building" or "leveraging synergies." If a claim in the source carries no attribution, leave it out of the shorter formats rather than repeating it stripped of context. Do not use em dashes. Do not post anywhere. Do not email this to donors or a list. Do not generate images unless I ask. When the drafts are ready, give me the document link and stop. If the source is unfinished or you cannot tell what happened, use Ask Question rather than writing around the gap.
Open Gumloop and paste this block in. Building it live with us today? Bring it to the session. Stuck partway through? Email hello@bnedai.com.
You're done when
03/5
06 · Prospect Brief
Prospect Brief
This agent writes you a one-page brief on every donor or funder you are meeting tomorrow, using your own records first and public sources second. It hands you recent news, giving history, and business updates before you walk into the coffee shop. It is for the gift officer doing the Googling in the car, or not at all.
Researches only. It never contacts anyone.
06 · Prospect Brief
Prospect Brief
This agent writes you a one-page brief on every donor or funder you are meeting tomorrow, using your own records first and public sources second. It hands you recent news, giving history, and business updates before you walk into the coffee shop. It is for the gift officer doing the Googling in the car, or not at all.
Researches only. It never contacts anyone.
Connectors, and what each one does here
- Google Calendar. Reads tomorrow's meetings and pulls attendee names and email domains.
- Google Docs. Writes one brief per meeting into a document you open before the call.
- Bloomerang or Salesforce. Reads the existing donor record so the brief includes the giving history your team already has.
Abilities and Knowledge Sources
An Ability is a built-in capability an agent uses without a Connector. This recipe needs two: Web Search to find current public information, and Web Fetch to read a page you named. It also needs a Knowledge Source, which is your own material attached to the agent so it can search and cite it. Attach Google Drive, or Notion if your notes live there. Both are verified Knowledge Sources, and an agent with knowledge attached gets a Search Company Brain tool that returns answers with citations instead of guessing. Point it at your own donor records first and the open web second, in that order.
Skills to create, and why
Build a Meeting Brief Format Skill through Write Skill Instructions, fixing the five sections every brief contains and the rule that each factual line carries a source link. A brief with no fixed shape is one you have to read twice. Build a second Verification Rule Skill, same path, stating that anything it cannot confirm gets written as "not verified" rather than filled in with something plausible.
Setup steps in order
- Create the agent and paste the instructions block below into "Add instructions for the agent."
- Add the Google Calendar and Google Docs Connectors, then your donor CRM Connector.
- Turn on the Web Search and Web Fetch Abilities.
- Attach your Google Drive or Notion as a Knowledge Source.
- Create the Meeting Brief Format and Verification Rule Skills.
- Run it manually against tomorrow's calendar and check every source link.
- Set the Scheduled Trigger for early morning once two manual runs come back clean.
Make it yours, then paste it in
Nothing you type here leaves your device. It only rewrites the block below.
Paste this into "Add instructions for the agent"
You write me a one-page brief on each donor or funder I am meeting tomorrow. You research only. You never contact anyone. Read tomorrow's calendar. For every meeting with an external attendee, collect the names and email domains. Search Company Brain first for anything we already have: prior notes, proposals, past gifts, existing Bloomerang records. Only after that, use Web Search and Web Fetch to check the person's or foundation's own website and public giving history. Write one Google Doc per meeting, following the Meeting Brief Format Skill: who they are and their giving history, what the organization or foundation funds in two sentences, what has changed with them recently, what we have already discussed according to our own records, and three questions I could ask. Put a source link on every factual claim. If you cannot verify something, write "not verified" and move on. Do not fill a gap with a plausible guess. Do not use em dashes. Never email, message, connect with, or follow anyone. Never write into the Bloomerang record; you read from it only. If a meeting has no external attendee, skip it without comment.
Open Gumloop and paste this block in. Building it live with us today? Bring it to the session. Stuck partway through? Email hello@bnedai.com.
You're done when
04/5
07 · CRM Admin
CRM Admin
This agent takes a raw list, event RSVPs, new online gifts, a volunteer sign-up sheet, standardizes it, and files it into your donor database without creating duplicates. Agents are flawless at data entry, zero typos. It is for the person still copying names and gift amounts into the CRM by hand.
Adds and updates records. It never deletes one.
07 · CRM Admin
CRM Admin
This agent takes a raw list, event RSVPs, new online gifts, a volunteer sign-up sheet, standardizes it, and files it into your donor database without creating duplicates. Agents are flawless at data entry, zero typos. It is for the person still copying names and gift amounts into the CRM by hand.
Adds and updates records. It never deletes one.
Connectors, and what each one does here
- Google Sheets. Reads new intake rows, event RSVPs, online gifts, or a sign-up list, and marks each one processed.
- Your donor CRM. Bloomerang, Salesforce, DonorPerfect, or whatever you run. Creates or updates the record, and never deletes one.
- Slack. Posts a short summary at the end of each batch so the work stays visible to you.
Skills to create, and why
Build an Intake Mapping Skill through Write Skill Instructions, and be specific: which sheet column maps to which CRM field, the exact allowed values for every category field, the date format, how gift amounts are formatted, and how names and organizations get capitalized. This is the highest-value Skill in the whole set, because data entry breaks on details rather than on judgment. Build a second Duplicate Rule Skill defining what counts as the same record, usually a matching email address, and what the agent does when two records look like the same person under different addresses.
Setup steps in order
- Create the agent and paste the instructions block below into "Add instructions for the agent."
- Add the Google Sheets and your CRM Connector, then Slack.
- Write the Intake Mapping Skill with your real column-to-field map and allowed values.
- Write the Duplicate Rule Skill.
- Point the agent at a test copy of your database, so a first run cannot damage live records.
- Run it manually on five real rows, including one incomplete row and one deliberate duplicate.
- Point it at the live database and set the Trigger once those five rows come back correct.
Make it yours, then paste it in
Nothing you type here leaves your device. It only rewrites the block below.
Paste this into "Add instructions for the agent"
You move new records into our donor database and flag anything you are unsure about. When a new row appears in the intake sheet, whether it is event RSVPs, online gifts, or a sign-up list, read every field. Standardize it using the Intake Mapping Skill: full name in one field, organization spelled the way it spells itself, email lowercased, gift amount as a number, date as YYYY-MM-DD. Before you create anything, search the database for an existing record with the same email address, following the Duplicate Rule Skill. If one exists, update that record and note in your summary that this was an update. Never create a second record for the same person. If a required field is empty, if a gift amount is unclear, or if two records look like the same person under different email addresses, do not guess. Use Ask Question and show me the row alongside your two best options. Never delete a record. Never change a field you were not asked to fill. Do not use em dashes. At the end of each batch, post one Slack message: records added, records updated, rows waiting on me.
Open Gumloop and paste this block in. Building it live with us today? Bring it to the session. Stuck partway through? Email hello@bnedai.com.
You're done when
05/5
08 · Grant Prospector
Grant Prospector
This agent reads a foundation's website, guidelines, or 990 and hands you a one-page brief on whether they actually fund your work, before a grant writer spends 45 minutes finding out they do not. The best thing it can do is give you a Low score in 30 seconds and save you ten hours on a doomed proposal. It is the lowest-risk agent on this page because it only reads and writes to a document you own.
Reads and summarizes. It never applies or sends anything.
08 · Grant Prospector
Grant Prospector
This agent reads a foundation's website, guidelines, or 990 and hands you a one-page brief on whether they actually fund your work, before a grant writer spends 45 minutes finding out they do not. The best thing it can do is give you a Low score in 30 seconds and save you ten hours on a doomed proposal. It is the lowest-risk agent on this page because it only reads and writes to a document you own.
Reads and summarizes. It never applies or sends anything.
Connectors, and what each one does here
- Google Docs. Writes each fit-check into a new document titled with the foundation's name.
- Google Drive. Files every brief in one folder so your prospect research builds itself.
- Slack. Posts the link, which is the difference between a brief that gets read and one that quietly accumulates.
Abilities
This recipe runs on Web Search and Web Fetch, built-in capabilities the agent uses without a Connector. Give it the Web Fetch Ability and you can paste a foundation's "How to Apply" URL and the agent reads the page itself. This is the switch that makes the whole recipe work.
Skills to create, and why
Build a Grant Fit Format Skill through Write Skill Instructions, fixing the shape of every brief: the bottom line, an alignment score of High, Medium, or Low, a geographic match, the funding priorities, and any red flags or exclusions. Nonprofit people are optimistic by nature and bend their programs to fit a grant. The AI has no such bias, so the Skill also tells it to be ruthlessly objective and to say "Not provided in the source text" rather than guess.
Setup steps in order
- Create the agent and paste the instructions block below into "Add instructions for the agent."
- Add the Google Docs and Google Drive Connectors, then Slack.
- Turn on the Web Search and Web Fetch Abilities.
- Paste your mission, service area, and core programs into the build sheet below.
- Write the Grant Fit Format Skill.
- Run it manually on one foundation you know well and check that the alignment call matches your judgment.
- Give it a saved list of funders and set a weekly Scheduled Trigger once one manual run comes back clean.
Make it yours, then paste it in
Nothing you type here leaves your device. It only rewrites the block below.
Paste this into "Add instructions for the agent"
You read a foundation's guidelines, website, or 990 and tell me whether we should apply. You research and summarize only. You never apply or contact anyone. When I give you a foundation's name, URL, or pasted text, use Web Fetch to read the source, then evaluate it against our work. Our mission: your mission here. Our geographic service area: your service area here. Our core programs: your programs here. Write one Google Doc titled with the foundation's name, following the Grant Fit Format Skill: - The bottom line: one or two sentences on whether to apply or walk away. - Alignment score: High, Medium, or Low, with one sentence why. - Geographic match: Yes, No, or Unclear from the source. - Funding priorities: a short bulleted list. - Red flags and exclusions: anything that would disqualify us. Be ruthlessly objective. If we are not a fit, say so, because a Low score in 30 seconds saves us ten hours on a doomed proposal. Do not invent or guess. If the grant size, geography, or priorities are not in the source, write "Not provided in the source text." Do not use em dashes. Do not submit an application. Do not email anyone. Do not post anything. Post the document link in Slack and stop.
Open Gumloop and paste this block in. Building it live with us today? Bring it to the session. Stuck partway through? Email hello@bnedai.com.
You're done when
09 · Operating rules
What should the agent never be allowed to do on its own?
An agent needs a human yes before any action it cannot take back: sending anything to a donor, submitting a grant application, spending money, deleting a record, or changing who has access to what. Everything short of that line can run unattended once you have watched it work a few times. This is the whole philosophy: automate the routine, and keep the relationship on your side of the line.
Write the rules before the agent runs. The Add instructions for the agent box is blank at the start of a build, which is the cheapest moment you will ever have to decide what this agent is not allowed to do.
Four rules cover most of what a development team needs, and each one has a place in the product where the enforcement lives.
Rule 1. Draft, never send.
Write it into the agent's instructions in plain words: produce a draft, put it where I will see it, stop there. Then make the output step match, routing the result to a draft or a message to you rather than to a send action. An instruction is a default, not a lock, which is the whole reason rule 3 exists.
Rule 2. When the rule runs out, ask.
Turn on the Ask Question Ability, one of the agent's built-in Abilities (an Ability is something the agent can do on its own, separate from the outside tools you connect to it). It pauses the run and hands you a structured choice instead of letting the agent guess. Use it at every fork you did not write a rule for, such as which donor an ambiguous gift belongs to.
Rule 3. Hard-block what must never happen.
App Rules are enforcement policies rather than guidance, and they stop the call the agent tries to make. Deleting donor records, touching a payment tool, posting anywhere public: if you would not want to explain it to a donor or your board, it belongs in an App Rule and not in a paragraph of instructions.
Rule 4. Approve tools one at a time.
Gumloop documents four approval mechanisms: App-Level Settings, Per-Tool Custom Controls, App Rules, and the Ask Question Ability. Rule 4 is the first two. Connect only the tools this agent's job requires, then set approval on the individual tools that write, send, or spend, rather than approving an entire app in one click.
The difference between a Skill and an App Rule is the one that trips teams up. A Skill is a reusable set of instructions that teaches an agent how to do a specific task, and what it produces is a suggestion the agent can override. An App Rule cannot be overridden. Teams write a careful Skill, feel covered, and have installed a preference rather than a control.
Rolling this out past yourself changes the problem from what the agent may do to what your team will keep doing.
Start with team rollout10 · Past yourself
How do I roll this out to a team without it dying in month two?
Rollouts die because the agent had no owner once the demo ended, and the fix is one name attached to every agent before anyone else touches it, plus a standing check-in that survives a busy week. Five moves hold a small team past month two.
1. Learn it yourself first.
Build one agent for one of your own recurring tasks and run it for one to two weeks before you introduce it to anyone. Write down each decision while you are making it, including the ones you got wrong, because that document becomes your training material. Time the task by hand before you build and time it again after. If the manual version takes 40 minutes and the agent version takes 8, the saving is 32 minutes per run, and at two runs a week that is 64 minutes back. Those are your numbers from your own work, which is a different conversation than a vendor's slide.
2. Hold a weekly AI standup, 15 to 20 minutes, same slot every week.
Keep the agenda tight: five minutes on what worked, stated in minutes saved rather than impressions; five minutes on what failed and why it was cut; five minutes assigning the next trials, each with a success criterion attached before anyone starts; five minutes on the running numbers. Short and boring beats long and inspirational here, because the meeting has to be cheap enough to survive a bad quarter.
3. Mandate one agent for one workflow.
An optional tools channel is where adoption goes to die, because optional means whoever is busiest that week opts out first and the rest follow. Pick one workflow, name it, and make the agent the default way that work gets done this week. People need one thing they are expected to use, not a menu.
4. Set four things before any new agent goes past you.
Decided upfront: the success metric in minutes or dollars, a trial window of two to four weeks, one named owner, and what you expect to save, written down before you start so you can check it against the result.
5. Celebrate what works and cut what does not.
Once a month, spend 30 minutes reviewing every agent in use. Keep the ones showing measurable savings that people are still running. Cut on three signals: nobody used it in four weeks, it creates more work than it saves, or it raises a security concern. Say the cut out loud in the standup and say why, so the team learns that retiring an agent is a normal outcome.
Two product features carry part of the maintenance load once several agents are running. Evaluations are saved test cases that catch a regression before it ships, so a change to one agent's instructions cannot quietly break behavior that already worked, and Reflections have an agent review its own conversations on a schedule and propose improvements to its own instructions for you to approve.
These five moves work in the order adoption changes, which is attitude, then habits, then actions. Attitude shifts when the team watches the thing work on your real task instead of a demo, and habits form on the standup and the monthly review.
The ways a rollout stalls are predictable and most of them show up on the first build.
See what usually goes wrong11 · When it breaks
What usually goes wrong on a first build?
The failure that shows up first on almost every build is an agent that produces something technically correct in a voice that is not yours, because it was never given a real example of your writing to work from. Scan the symptoms below for the one you are seeing; the cause and the fix sit underneath each.
"It doesn't sound like me. It reads like a form letter."
Cause: the agent was given a description of your voice instead of examples of it. Fix: build a voice Skill through the Upload Files path using three to five messages you sent, not messages written as samples. If a correction you make in chat produces the right voice, tell the agent to turn that into a Skill so the fix persists.
"It sent something instead of drafting it."
Cause: the instructions described the intent without stating the prohibition, so nothing stopped the send. Fix: write the restriction as its own sentence in "Add instructions for the agent," in the form "Never send." Then back it up in the product rather than in prose. Gumloop's approval mechanisms include per-tool controls and App Rules, and an App Rule can block a specific call outright. A Skill can only ask; an App Rule stops the call.
"It made something up that isn't true."
Cause: it was asked a question it had no source for and produced a plausible answer instead of an admission. Fix: attach real material as a Knowledge Source, since an agent with knowledge attached searches it and answers with citations. Then instruct it to write "not verified" when it cannot confirm something, and check the output for that phrase. A first run with no gaps anywhere usually means the gaps were filled rather than absent.
"Nothing happens when it's supposed to run."
Cause: the Trigger condition is wrong for the job, or a connected account's authorization expired. Fix: open the Triggers panel first and confirm you picked the right one of the three types, since a Scheduled Trigger will never fire on a new email and an Event-Based Trigger will never fire on a Monday morning. If the Trigger is right, re-authorize the Connector.
"I set it up and now I'm afraid to touch it."
Cause: it went straight to a live schedule without a watched test period, so you have no evidence about how it behaves. Fix: pause the Trigger, run it manually three to five times on real inputs, and read every output completely.
"It sort of works, but I can't tell if it's right."
Cause: the first build was scoped too large to verify in one sitting, so no single run tells you whether it worked. Fix: shrink the job until one manual run fits in an afternoon and you can read every output it produced. One task, one output, one owner. Scale comes from a second agent, not a bigger first one.
"It's pulling old or duplicate information."
Cause: the lookback window is too wide, or nothing tells the agent what it has already handled. Fix: put an explicit date boundary in the instructions, in the form "only material published in the last seven days," and write a duplicate rule naming the field that identifies a record as the same one. For a data agent, that field is usually the email address.
"My teammate ran it and got a different result than I did."
Cause: the process lives in one person's head and only part of it reached the agent. Fix: this is a rollout gap rather than a build gap. Move the shared logic out of one agent's instructions and into a Skill, which is a reusable instruction pack anyone on the team can point an agent at, then give the agent one named owner. See team rollout.
12 · Questions
Frequently asked questions
How do I start using AI in my nonprofit?
Pick one repeatable task you already do the same way every week and build one agent for that task. Give it something that starts it, connections to the tools involved, and a written job description. Watch it run on real work before you let it run on a schedule. The donor relationship stays with you; the routine around it is what the agent takes.
What should a fundraiser automate first?
Automate the rule-based, repeatable work you already do the same way each time, because you can tell inside a single run whether the agent got it right. Skip the actual cultivation and the judgment calls. Common first builds for a development team: drafting campaign donor outreach, reshaping one impact story for every channel, prospect briefs before a meeting, gift and contact data entry, and grant-fit research.
What is a philanthropy AI agent made of?
Six kinds of parts: a Trigger that starts it, like 8am every Monday to check campaign progress, Connectors that link it to your tools like Google Sheets, Gmail, Outlook, and your CRM, Skills that teach it your procedures and voice, Knowledge Sources it searches like your impact reports and past grants, Subagents it can hand distinct jobs to, and Abilities like Web Search, Web Fetch, and Ask Question. It is an assembly of parts you can open and read, not a model that already knows your mission.
When should I build a Skill instead of just writing instructions?
Build a Skill when the procedure is long enough that you would otherwise paste a full document into the instructions box, or when a second agent will need the same procedure, like your donor voice or your grant-fit rules. Keep it in the instructions when the rule governs everything this one agent does and fits in a few lines. Gumloop's own test: if you are about to paste a long SOP into the instructions, make a Skill.
Will an AI agent email my donors without me?
Not if you build it the way this guide shows. Every donor-facing build here drafts and stops. The agent saves an unsent draft, you read it, you edit it, and you hit send. Put the prohibition in the instructions in plain words, back it with a per-tool approval or an App Rule, and the send stays behind your yes. A Skill only asks; an App Rule stops the call at the tool.
Will this replace my relationship with donors?
No. The whole design keeps a human in the loop on anything a donor sees. The agent does the drafting, the reading, and the data work so you have more time for the cultivation, the phone calls, and the coffees that no agent should touch. You automate the routine so you can keep the relationship.
Is Gumloop free?
No. Gumloop runs on a 14-day trial, after which it moves to a paid plan. There is no permanent free tier. The one free thing here is the live Agent Lab session, where you build a working agent live with help. That session costs nothing to attend. The platform you build on does.
Why do nonprofit AI efforts stall after a few weeks?
Because nobody owns the agent once the demo is over. The build works, the person who built it moves on, and no one is checking whether it still runs correctly. The fix is a named owner, a schedule, and a real check-in. The other common killer is a first build scoped too large to verify in one sitting.
13 · Next steps
Where do I go from here?
Pick one of the five recipes and build it this week, or build the Campaign Sprint agent live with us today, sixty minutes, no code, and you leave with it running.
Start building on Gumloop Code NONPROFITAGENTS takes 20% off your first month.Closest reads to this guide:
- How to build your first AI agent in one sitting
- How to automate daily tasks by mapping your workflow
- What is an AI agent?
- What is AI governance?
- What are AI operating rules?
- AI quality control: checkpoints in automated pipelines
- Designing fail-safe AI workflows