Capacity
without
collapse.
Practical AI workflows for nonprofits navigating transition, designed for steadiness, not speed.
Robinson, M.S.
Nonprofit strategic advisory partner
You are
already
doing the work.
This is a peer-to-peer working session, built for leaders who are already carrying too much, and looking for ways to set some of it down without dropping the mission.
- 01Name where capacity is being drained.
- 02Map your Requests & Follow-Up Capacity Flow.
- 03Stabilize one workflow with AI as infrastructure.
- 04Leave with a plan you can apply this week.
I'm Latoya.
I've sat where you're
sitting.
I am a longtime nonprofit Executive Director, the founder of LUR Growth, and a strategic advisory partner to leaders and boards designing capacity-building infrastructure that lasts.
In the nonprofit sector.
Programs, fundraising, operations, and the late-night work in between.
As an Executive Director.
Carrying the inbox, the funder relationships, and the institutional memory.
Workflow design & implementation.
Practical, low-jargon systems leaders can actually maintain.
Responsible AI as infrastructure.
Reducing pressure on people; protecting funder and community trust.
Today's agenda
is your
pre-work.
I read every word. Names and organizations stay anonymous — but here is the room, in your own language.
What you're holding alone
- →Daily financial decisions & grant balances
- →The one system login no one else has access to
- →Board prep — a few hours every month
- →Staff supervision + strategy + an event, at once
- →Program details that "depend on me"
Where workflows break
- →Follow-up
- →The middle of onboarding
- →Going back through old emails to find what was said
- →"It depends on me"
- →Standing up a brand-new tool the team has to learn
What you want from today
- →A clear picture of what I can hand off
- →Tools that handle the redundant tasks
- →Less stress that something will drop
- →To understand AI's role in our work
- →Time back for the bigger picture
That's the room. That's today's agenda.
The sector has a
systems
problem.
Funding is uncertain. Staff is shifting. Accountability is rising. And most of us are still running strategy, follow-up, and continuity on people alone.
Three pressures, all at once.
Funding uncertainty.
Grant cycles compress. Renewals slip. Multi-year commitments quietly become single-year. Cash flow is harder to forecast than at any point in the last decade.
Staffing shifts.
Tenured staff leave. New hires take six months to ramp. The person who held the spreadsheet, the relationships, the inbox, gone, often without a replacement.
Rising accountability.
Funders want clearer outcomes. Boards want sharper reporting. Communities want faster response. The bar is rising while the bench is shrinking.
When the mission depends on one person remembering everything, the mission is one bad week away from breaking.
This workshop is
not
about efficiency.
We are not here to make you faster, busier, or more "productive." We are not here to add another dashboard or another tool.
Which pressure is
hitting your org
hardest
right now?
Before we look at the drain inside the building, let's name the weather outside.
- A Pick one of the three pressures: funding, staffing, or accountability.
- B Name the moment in the last 30 days where you felt it most.
- C Ask each other:"If this stays the same for six more months, what breaks?"
No fixing. No advising. Just naming. We'll come back to it.
Most of the loss is
invisible.
Burnout is not caused by the work you can see. It's caused by the work nobody assigned, nobody tracked, and nobody noticed, until it didn't get done.
Where the capacity actually goes.
Not bad people. Not lazy teams. Patterns, recurring shapes that show up in nearly every nonprofit during transition.
Requests pile up.
Every channel, email, text, hallway, board chat, feeds the same one inbox: yours.
Decisions loop.
The same question gets reopened in three meetings because no one wrote down the answer.
Follow-up vanishes.
Action items live in someone's head, not in a system. They get done, or they don't.
Continuity rests on one person.
One leader holds the whole map. When they're sick, on vacation, or gone, the map goes with them.
The work that
holds
the org together.
None of these show up on a job description. All of them show up when they stop happening.
You can probably name three more before I finish this sentence.
Where is your
capacity
being
drained
this week?
- A Name three things you're personally holding that nobody else in your org could pick up tomorrow.
- B Circle one that, if dropped, would cause real harm.
- C Mark the channel it lives in: email, head, calendar, sticky note, slack, nowhere.
"What's the one I named, and where does it actually live?"
The Requests &
Follow-Up
Capacity Flow.
A simple, three-stage map of how work enters your organization, how decisions get made (or don't), and how follow-up gets closed (or vanishes).
How work moves through your organization.
Five questions. One workflow at a time.
Pick one workflow you know is leaking, donor follow-up, hiring, board prep, and walk it through these five questions. The drawing is the work.
Where does it enter?
Email, text, hallway, calendar, form. Name every channel, even the awkward ones.
Who sorts it first?
If the answer is "me, in my head," circle it. That is the drain.
Who decides, by when?
A real name and a real date. "Soon" is not a date. "We" is not a name.
Where is it written?
If a future you couldn't find it in 30 seconds, it isn't written down.
How does it close?
A confirmation back to the requester. Otherwise the loop never closes, even if the work got done.
From reactive operations
to designed infrastructure.
One of you is hiring this week. This one is for you.
Onboarding lives in the founder's head.
- →New hire shows up Monday. Orientation covers what comes to mind that day.
- →Week two, someone realizes they were never shown the shared drive.
- →Founder answers the same five questions for six weeks straight.
- →The 30-60-90 conversation never happens because there's no map.
- →The new hire's "ramp" depends on the founder's calendar that month.
Onboarding lives in a system.
- →A standing onboarding doc with the five things every new hire needs in week one.
- →AI drafts a welcome packet from the job description and team norms — founder edits in 10 minutes.
- →A 30-60-90 template, pre-filled, ready for the first one-on-one.
- →Founder's calendar isn't the ramp. The system is the ramp.
- →The next hire doesn't start from scratch — they inherit the doc.
Map your
own
Capacity Flow.
Pick the workflow you circled earlier. We'll walk it through Intake → Decision → Follow-Up together.
Sketch the three boxes. Label them Intake, Decision, Follow-Up.
For each box, write where it actually lives today, and who holds it.
Mark with a star ★ the one box where this workflow most often breaks.
We'll come back to your starred box in the next section.
For care.
Not
for efficiency theater.
Used responsibly, AI is a quiet structural support: it remembers, drafts, and surfaces, so the people in your organization can do the work only people can do.
Before we go further —
what is
this thing?
If you've never opened one of these tools before — and several of you wrote me to say you haven't — this is for you. No catching up required.
It's a website you log into.
chatgpt.com · claude.ai. You type, it answers. No download, no install. Free tier on both.
It writes back in plain English.
Ask it the way you'd ask a smart colleague. "Help me draft a thank-you to a donor who gave us $500 last spring."
It's a draft, not a send.
You read it. You change what's wrong. You decide whether it leaves the building. That's the whole job.
If you only do one thing after today: open one of the two URLs above and ask it to help you with the single workflow you mapped in the last activity.
What this looks like
in my week.
I'm not asking you to do anything I'm not already doing, with the same guardrails, on real client work.
Inside the building, on the hard weeks.
- →Grant calendar & reporting. Deadline tracking, first-draft narratives pulled from program notes for me to verify.
- →Board prep. Pre-meeting briefs, post-meeting decision logs — nothing reopened next month.
- →Donor stewardship. Thank-you drafts informed by giving history. I sign and send.
- →Staff & volunteer follow-up. Commitments tracked across channels so nothing lives only in my head.
Quiet scaffolding behind a one-person shop.
- →Inbox triage. AI sorts client requests, surfaces what actually needs me, and drafts the rest.
- →Meeting prep & recap. Briefs before, summaries after — every commitment captured.
- →First drafts in my voice. Proposals, frameworks, follow-ups — always edited by me before they leave.
- →Knowledge memory. A private library of my past work I can search like a colleague who never forgets.
Built with you, never bolted on.
- →We map first. No tools until the Capacity Flow is on paper and the broken stage is named.
- →One workflow at a time. Donor follow-up, or board prep, or hiring — not all at once.
- →Humans always send. AI drafts, summarizes, reminds. Staff edits, approves, and keeps the relationship.
- →Written guardrails. A one-page AI use policy your board, staff, and funders can all read.
What AI should and should not do.
| In your nonprofit | Yes · AI as infrastructure | No · AI as replacement |
|---|---|---|
| Donor relationships | Draft a thank-you informed by past gifts and notes, for a human to edit. | Auto-send a thank-you with no human read of the tone or context. |
| Meeting prep | Summarize a long thread, surface decisions and open questions for the agenda. | Replace the actual conversation with an AI-generated decision. |
| Email follow-up | Track which threads are unanswered, draft check-ins, surface stuck ones. | Send program responses to community members without staff review. |
| Reporting | Pull program data into a draft funder report, for the ED to verify and finalize. | Submit auto-generated impact numbers without human reconciliation. |
| Sensitive cases | Flag a complex situation. Stay out of the response. | Generate guidance for survivors, clients, or vulnerable communities. |
Small. Specific. Reversible.
Each one reduces tool overload, protects continuity, and is safe for non-tech-savvy teams to adopt.
Every commitment, in one place.
Scans recent meetings and emails for open "I'll get back to you" promises — surfaces them in a weekly digest you actually read.
For the person who goes back through old emails to find what was promised.
A 2-page brief in two minutes.
Before each meeting, AI assembles history, last commitments, and three questions worth asking — drafted, not sent.
For the person spending hours every month on board meeting prep.
A starting point, never the send.
For routine asks, AI gives a respectful first draft in your voice. You edit the human parts and send.
For the leader kept up by "tasks not getting done correctly or completely."
Nothing slips. Nothing feels canned.
Each gift triggers a personalized draft from the donor's history and notes. You review, edit, send within 48 hours.
For when stewardship rests on one person — and one bad week breaks the cadence.
Let me actually
do
one of these,
in front of you.
I'll take the workflow this room cares about most, open the tool, and run the starter prompt — live, unedited, mistakes and all. You watch the rhythm: ask → read → edit → you decide.
- 01A real example — pulled from your pre-work, not a demo script.
- 02The prompt, typed in plain English — the same one on the workflow card.
- 03The first draft — and me deciding what's wrong with it.
- 04Where the human stays in — tone, judgment, the send button.
If we're tight on time, this is the first thing I skip — and you already have every prompt you need to do it yourself.
You don't have to build it yourself.
At LUR Growth I build custom infrastructure for clients who want it — but you don't have to. One well-chosen system, set up cleanly, can carry you a long way.
Bloomerang
Donor CRM, email marketing, online giving, retention scoring. Priced by record count. Volunteer & Giving Platform sold separately.
Little Green Light
Transparent published pricing. Donor records, mailings, gift entry. Less polish, fewer bells & whistles, much lower overhead.
Keela / Aplos
Newer all-in-ones with AI assists for emails and reports baked in. Pricing tiered by contacts, in line with Bloomerang.
Subscribe to a system
- Live in days, not months
- Vendor handles updates & security
- Real support & user community
- Clean funder & audit story
- Recurring monthly cost
- Pricing rises with your records
- Workflows fit their model, not yours
- Add-ons quietly stack up
Build something internal
- Fits exactly how you work
- No per-record pricing tiers
- You own the data, fully
- AI lowers the build cost dramatically
- Someone has to maintain it
- Security & backups are on you
- Funders may ask harder questions
- Bus-factor risk if the builder leaves
My honest take: I build my own — that's the work at lurgrowth.com. But a clean off-the-shelf tool is a real, valid choice. The wrong move is staying stuck between the two.
One tool.
Used well.
You don't need a stack. You need one thinking partner you trust — and the discipline to actually open it.
Claude Chat
The one I open every morning. Drafting, summaries, board prep, the hard email you've been avoiding. Ask it the way you'd ask a sharp colleague — in plain English.
Claude Projects
A private workspace per topic — one for the board, one for a grant, one for a program. Load the context once; it remembers it every time you come back.
These are real and they're powerful — and they're a Part 2, not today. When your Capacity Flow map shows you actually need one of them, that's the conversation. I'm building one of these in the open right now (Hermes Agent) so you can watch how the sausage gets made before you'd ever adopt it.
You don't need to be
tech-savvy
to do this responsibly.
You need clear rules, a small set of tools, and the steady habit of keeping people in the loop where it matters.
A human always sends.
No external message, to a donor, funder, partner, or community member, leaves without a person reading it.
No personal data in tools you don't trust.
Donor records, client cases, HR, only in vetted, contracted tools. We'll show you the short list.
Disclose where it matters.
Funders increasingly ask. A one-line policy in your operations doc protects trust and gets ahead of the question.
Fewer tools, integrated well.
One CRM, one calendar, one inbox, one doc system, connected. Not seven things AI tries to bridge.
Pick your
one
AI workflow.
And one guardrail.
We've seen the four. Now choose the one that matches the box you starred. Small. Specific. Reversible.
Circle one: Follow-Up Tracker · Meeting Prep · Email Drafting · Donor Acknowledgment.
Name the human in the loop. Who sends, who reviews, who keeps the relationship.
Write your one-line guardrail. The thing that stays human, no matter what.
Then share with your neighbor. One sentence each.
Your first prompt.
You picked a workflow and a guardrail. Here's the exact thing to paste — fill the brackets, send it tonight.
Here's the raw material: [paste your notes, emails, or the messy version].
Help me with a first draft. Keep it [warm / brief / plain]. I'll edit before anything leaves the building — the part that stays human is [your guardrail].
Tell it what's wrong in plain words: "too formal," "you missed X," "shorter." It revises. That's the loop.
Once is the win. The habit is next month's problem. Today, just get one draft you didn't start from scratch.
A workspace that remembers, a system that runs it, a custom build — that's the engagement. That's where I come in.
Start
small.
Start
this week.
One workflow. One owner. One closed loop. That's enough to begin.
Refine your Capacity Flow map.
Spend 30 minutes finishing the diagram you started here. Show it to one other person on your team.
Your pre-work already named the workflow — don't restart. Finish the one you sent me.
Stabilize the starred box.
Give the broken stage one home, one owner, and one written-down standard. AI can wait, the system is the win.
Add one AI-assisted workflow.
Pick the lowest-stakes one from today's four. Give it 30 days, with a human always in the loop.
The mission should not depend on one person
remembering everything. It should rest on
infrastructure that holds, even on the hard weeks.
Let's keep
building
capacity
that holds.
Supporting leadership and boards in designing and activating capacity-building infrastructure that protects people, strategy, and mission continuity.
Robinson, M.S.
Nonprofit strategic advisory partner