Notes from the field, every week.

Hi friends.

Last week I wrote a short post about spring cleaning. A few weeks of using Claude and ChatGPT Codex to clean up my files, my inbox, and the dusty corners of my Google Drive. Not to add new tools. Just to use the ones I already pay for, the way they were built to be used.

It did not stop there. The same instinct followed me home. I cleaned out my closet. Then the basement. Then back to the screen, where I started pulling the Anyone Can Fundraise website apart the same way. More on that one in a future edition, but the short version is that the site has been quietly out of step with how the work actually looks now, and it is time to fix that. Same approach as the inbox. What is already there. What can be cleaned, sharpened, made to work properly, before anything new gets added.

What surprised me about the original post was the response. People sent me notes about their own overflowing inboxes, their own unsearchable shared drives, their own untouched features. The same pattern almost everywhere. Capacity hiding inside the systems we already own. Tagging rules nobody set up. Folder structures nobody finished. Automations sitting there the whole time. Closets, basements, websites, drives. The pattern does not change.

This week I am taking that conversation off the screen. Two rooms. Two very different audiences. Same thread running through both.

Tuesday, May 26: Toronto Tech Week

Jonathan Hutchinson is hosting a Toronto Tech Week event on Tuesday evening at the Christie Ossington Neighbourhood Centre, and he has been generous enough to invite me to talk through what is actually working right now for nonprofits using AI. Practical, hands on, no hype. The same pattern from the spring cleaning piece, but live, with people in the room asking the hard questions.

There will also be a pop up featuring Angela Moritsugu's art.

A few free spots are still open: luma.com/y1j68dq1

If you are in Toronto, come. If you know a nonprofit leader who should be in that room, send them the link. The whole point of the night is the conversation, and the conversation is better the more practitioners we have in it.

August 6 and 7: Decoupled Days, Montréal

This one is the longer format conversation. Preston So invited me to lead a workshop at the ninth edition of Decoupled Days, the premier nonprofit headless conference, August 6 and 7 in Montréal.

My workshop is called "Your nonprofit already has everything it needs; it just cannot think yet."

It walks through how Anyone Can Fundraise connects existing nonprofit tools, CRMs like HubSpot and Salesforce, Google Workspace, email platforms, donor databases, with AI through MCP (Model Context Protocol) servers, Claude, ChatGPT, Microsoft Copilot Studio, and modular skill architectures.

No rip and replace. No six month implementation. Just wiring intelligence into the systems organizations already use and trust.

Attendees will see real examples from production. A donor reactivation workflow that pulls from a live CRM. An outreach system that researches prospects and drafts personalized emails. A grant discovery tool that matches opportunities to an organization's mission. All built on top of tools the nonprofit was already paying for.

This is not just a technical story. It is about making AI accessible to organizations that do not have engineering teams, that cannot afford enterprise contracts, and that serve communities most technology vendors overlook. What actually gets nonprofits to adopt AI, versus what just gets a wow in a demo.

Early registration is open this weekend only at the lowest price, CAD 35 for the talk plus the rest of the conference.

Why both rooms matter

Toronto is the short room. Tuesday night. Local nonprofits, capacity builders, technology people who serve communities most vendors overlook.

Montréal is the long room. Two days. Practitioners building the connective tissue between AI and the systems organizations already run on.

The conversation is the same in both. Stop chasing the next platform. Look at what you already have. Make it think.

If we are going to make AI useful for this sector, it will not be because someone shipped a new app. It will be because we finally learned to use the ones we already paid for.

Hope to see some of you Tuesday. And in Montréal in August.

Best, Micah

Eleven days from now, on Tuesday May 26, I am co-hosting a free workshop in Toronto with Jonathan Hutchinson as part of Toronto Tech Week. Jonathan runs IT at Christie Ossington Neighbourhood Centre and has spent years defending small organizations against real threats with limited resources. I will be running a live workshop on using AI tools for fundraising. Hands-on demos with Microsoft 365 Copilot, Salesforce, Google Workspace, and Claude. Practical, not a lecture. The kind of evening where you walk out having actually seen what is possible inside your own stack.

Tuesday, May 26, 6:45 to 9:15 PM. George Chuvalo Neighbourhood Centre, 50 Sousa Mendes Street, Toronto. Free. Light refreshments. Open to staff at nonprofits, charities, shelters, and small businesses. Plenty of structured networking time too.

Save your spot here: https://luma.com/y1j68dq1

Why does this feel urgent right now? Because of the week I just had. The riskiest thing a nonprofit can do right now is adopt a new system. The most useful thing it can do is connect the ones it already has. That has been the through line of every conversation, every post, every demo this week. And it is the conversation we are going to have on May 26.

Also from this week:

The subway moment. Heading home from BuildGood, twenty minutes left on the ride. I opened a tool on my phone and knocked out follow-up emails to the new connections I had made, a full conference report, and a head start on some client work, all before I got off the train. The part that mattered was not the tool. It was getting my evening back to spend with my family instead of catching up at 11pm.

At BuildGood. A short note from day one. Year two of the conference, still a bit rainy, more people than last year.

Why bother with the cutting edge? A development director stopped me mid-call last week and asked the question I should have been ready for. My first answer was bad. Here is the answer I wish I had given him, and why I think the gap between fundraisers who learn this now and fundraisers who wait is going to be the most consequential thing in our sector over the next eighteen months.

The riskiest thing a nonprofit can do right now is adopt a new system. The whole sector has been trained to believe the opposite. Migrate to this platform. Replace that CRM. Here is what I am seeing instead, and why connecting the tools you already have is changing the conversation.

Mindful Moment. A small thing I built one morning. A 60-second meditation right inside Salesforce, on every contact record. About 20 minutes of work in Claude Code. Before a fundraiser hops on a donor call, they can take a beat to center.

The pattern matters more than the platform. That is the thread running through every one of these posts, and it is exactly what we will be working through together on May 26. See you in Toronto.

I had a long back and forth with Claude this week. Client work. Lots of second-guessing. The imposter voice doing its usual thing, telling me the ask was too big, that I should pull back, soften, hedge.

A few days later, after the client said yes, Claude pulled the receipts.

"This is the result. The email you almost second-guessed just doubled your asked hours through the end of the engagement. Worth noting for the next time the imposter voice tells you a strong piece of work is too much. It was not."

I read it twice.

Not because the project was the headline. The headline was that Claude remembered. It knew who I am. It knew what I had been wrestling with. It knew the specific shape of my self-doubt. And it served the receipts back when the outcome came in.

That is new. And honestly, it was a little hard to take in the moment.

But that is the work. AI with memory is not a feature. It is a different kind of relationship. A collaborator who pays attention. A peer who is willing to tell you the truth, kindly, when the imposter voice gets loud and the data is on your side.

For nonprofit teams, this is the part that matters. The platforms are already there. The data is already there. The people who know the work are already there. The missing piece has been the thread connecting all of it. That is what AI is becoming now.

Also from this week:

Resilience. A short note for nonprofit leaders, charities, and small business owners about what it actually takes to keep showing up when the conditions are not making it easy. The tools are there. The people are there. The mission is real.

Prospect Finder dashboard, live inside Claude. A quick demo of a tool I built that pulls real data from HubSpot, Gmail, Calendar, Salesforce, Blackbaud, or none of the above, tracks who you have reached out to, and shows you what came of it. Anyone can do this. That is not a marketing line. It is literally the thing.

Coming up: Decoupled Days, Montreal, August 6 and 7. I will be running a workshop at Decoupled Days this summer called "Your nonprofit already has everything it needs; it just can't think yet." The platforms, the data, and the people are already in your stack. AI is the connective tissue that turns all of it into actual workflow.

Because the mission does not wait.

I am ending this week on something that matters more to me than any tool I have built.

Earlier this week I wrote about a conversation with a nonprofit leader who runs a small team. Two full-time people. Everyone else part-time. The ones whose work expands to fill whatever needs doing. I asked the question that changed everything for me: are we working the correct hours? Not more efficient hours. The correct ones.

That post landed with people in a way I did not expect. So I want to share what came out of it.

I built two AI assistants around this one idea. No surveys. No scores. No wellness questionnaires. Just conversations.

The first is for the people on your team. It works like a thoughtful colleague who asks: did the time you saved actually come back to you, or did it quietly become more work? You describe your week, and it reflects back what it notices. "You said AI saved you three hours on grant research. And then you said you used that time to start a new project. What if you had just gone home?"

The second is for leaders. The manager. The ED. The person carrying the weight of everyone else's wellbeing who rarely gets asked about their own. It starts with a simple question most people never hear at work: how are you, actually? Then it gets to the harder one: when your team gets time back from these tools, what is happening to that time?

Here is what makes me hopeful. Every leader I have talked to about this cares about the answer. They want their people to work less, not just differently. They just never had a framework for making that happen. That is what these tools are for.

A grant writer's research drops from three hours to twenty minutes. That is two hours and forty minutes back. The right answer is not more grants. The right answer is a shorter day.

Both tools work in Claude, ChatGPT, Google Gemini, and Microsoft Copilot. They were built for nonprofits because this is where I see the pattern most, but the question applies everywhere: if your team is more productive and nobody is working less, something got lost along the way.

It does not have to stay lost.

If this sounds like your team, I would love to figure it out together. Happy Friday. Close the laptop.

Yesterday I watched someone build an AI tool that does in minutes what used to take her days.

She is not an engineer. She is a nonprofit professional who needed media contacts in a dozen new counties where she had zero. Finding five contacts used to take a couple of hours. Finding twenty in new territory would take days.

So she studied how an existing tool worked, wrote out exactly what she needed, and asked AI to build it for her. Then she iterated. Adjusted the categories, refined the output, tested it on real work. Multiple rounds. Line by line.

That is the shift I keep seeing. People go from using tools someone else built to building tools that solve problems only they can see. Because nobody else has their job. The person closest to the problem is always the best person to define the solution.

Yesterday we ran a session where the whole point was that transition. From consumer to creator. Not writing code. Writing instructions. If you can explain how you do something to a new hire on their first day, you can build your own AI skill.

One person built a media contact finder. Another customized an accountability tool that knows their specific blind spots, not generic advice. Someone else started sketching out an event planning checklist that catches the things they always forget.

None of these are technically complex. All of them are specific. That is the difference.

The three rules we kept coming back to: start from something that already works, be specific about what you need, and test it on real work. If it does not save you time on day one, keep iterating until it does.

This is the part of capacity building that sticks. Not a training session people forget by Friday. A tool someone built on Wednesday that they are still using next month.

Anyone Can Fundraise AI Weekly - Edition 22

This week we crossed a line I have been building toward for months. We ran our first Build Your Own AI Skill session, and watched nonprofit professionals go from using tools we gave them to creating tools that solve problems only they can see. One person built a media contact finder for a regional expansion. Others started customizing accountability tools tuned to their own blind spots. No code. No engineering background. Just written instructions and real iteration.

The theme across this week's posts is the same: the organizations making progress are the ones building inside their own workflows, not waiting for someone else to hand them the answer.

The Three Questions. What real AI adoption looks like: from sorting a list, to pulling a report, to building your own assistant. It happens in about a month if someone is there to answer the first question.

Build Where Your Team Already Works. Why the best AI tool is the one that lives inside the ecosystem your team already uses. Google Workspace, Microsoft 365, wherever they are.

What Stays After the Grant Ends. Three quarters of nonprofits say their funders do not understand their AI needs. The capacity building that matters is the kind that changes Tuesday morning, not the kind that lives in a binder.

It's Friday. And for a lot of you, March 31 is four days away.

Which means this week probably looked like: final appeal pushes, last-minute gift processing, board reports due yesterday, and that donor who said they'd confirm by March and still hasn't called back.

I get it. This is the sprint at the end of the marathon.

Before you log off for the weekend, here's what dropped in Edition 21 of Anyone Can Fundraise AI Weekly:

This week I wrote about the broken phone systems we've all been stuck in. You call a nonprofit, a condo office, a service provider, and you get the same dead-end voice menu. The real opportunity with AI isn't putting a smarter robot on the front line. It's freeing your people from their screens so they can actually pick up the phone and be human.

I also announced that I'll be presenting a live AI fundraising workshop at Toronto Tech Week on May 26th alongside Jonathan Hutchinson, who runs IT at Christie Ossington Neighbourhood Centre. Cybersecurity, AI-powered fundraising workflows, and how small to midsized organizations can punch above their weight with the tools they already have. Free event, 6:45 to 9:15 PM at George Chuvalo Neighbourhood Centre.

Save your spot for the event: https://luma.com/y1j68dq1

Subscribe to Anyone Can Fundraise AI Weekly so you never miss an edition.

Now close the laptop. Don't check the CRM. Whatever didn't get done by 5 PM today will still be there, and you'll handle it better after some rest.

Have a great weekend. You earned it.

We keep hearing about the "AI maturity model." You know the one. Start with human-done work. Slowly automate pieces. Move through stages until you reach the promised land of fully autonomous AI agents.

It sounds logical. It sounds inevitable. And I think it is fundamentally wrong, for nonprofits especially, but honestly for everyone.

Here is the problem. These frameworks treat humans as the bottleneck and AI as the destination. They assume the goal is to remove the human from the equation, one task at a time, until the machine runs the show. For the for-profit world, this is already producing work that is faster and cheaper but noticeably worse. Efficient? Sure. But stripped of the judgment, empathy, and creativity that made it valuable in the first place.

And let's talk about what is actually happening. Organizations are using these frameworks to justify cutting headcount. They are firing people and calling it transformation. But removing the humans does not improve the end product. It hollows it out. Because everything we produce is ultimately for humans, and AI alone cannot deliver the nuance, empathy, and creativity that humans bring to that work.

For nonprofits, it is even more dangerous. Our work is fundamentally human. We are asking people to give their money, their time, their trust to a cause. That requires connection. It requires a person who understands why a donor gives, what moves a community, and how to tell a story that makes someone feel something.

But here is what I am not saying: I am not saying ignore AI.

The real opportunity is not automation. It is synergy. AI as a collaborator, not a replacement. A tool that makes the people you already have better, faster, more informed, so the work they produce is not just more efficient but genuinely superior to what either could do alone.

When a fundraiser uses AI to research a prospect and then writes a deeply personal outreach message informed by that research, that is better than either could produce solo. When AI helps analyze your donor data so you can have a more meaningful conversation with a major gift prospect, that is the future.

The organizations that will thrive, nonprofit and for-profit alike, are the ones that stop cutting people and start empowering them. Stop chasing full automation and start building real human-AI collaboration into everything they do.

Stop following the maturity model. Start building the synergy model.

In case you missed it this week:

"So you help people use their computers better?" My kid summed up my entire career in one sentence at dinner. And he was not wrong.

The most important person in your AI meeting has their arms crossed. Every nonprofit has someone who survived the last "transformational" technology rollout. They are not your obstacle. They are your most important stakeholder.

This is the 20th Edition of the Anyone Can Fundraise AI Weekly.

This week I published two posts that hit a nerve, and they're connected by the same thread: nonprofits are using AI, but most aren't getting anywhere with it.

The Capacity Post: I wrote about how the real question isn't whether AI exists. It's whether we're equipping nonprofits with the skills and frameworks to use these tools intentionally, with the platforms they already have. Not a new system. Not a six-month implementation. The tools sitting in your Microsoft 365 or Google Workspace right now.

The Prompting Trap: The 2026 Nonprofit AI Adoption Report dropped and the numbers are stark. 92% of nonprofits are using AI. Only 7% say it's meaningfully changed their work. The gap? Most teams are stuck in one-off prompting mode, no memory, no context, no connection between Monday's work and Friday's.

Main Story: What AI With Memory Actually Looks Like

I want to make something concrete that I kept abstract in those posts.

Imagine your development director sits down on a Monday morning. She doesn't open ChatGPT and type "write me a follow-up email to a major donor." She opens an AI assistant that already knows her portfolio. It knows that David Chen gave $25,000 last year at the gala, that he mentioned his daughter's involvement in youth mentorship programs during a site visit in October, and that the last touchpoint was a thank-you call in January.

The assistant drafts a follow-up that references the mentorship connection, flags that David's giving anniversary is coming up in six weeks, and suggests a personalized invitation to an upcoming program tour.

She didn't prompt any of that. The system remembered.

Now imagine the grants manager opens the same system. She's not starting from scratch on a new proposal. The assistant already knows the organization's mission statement, last year's outcomes data, the language that worked in the three grants that got funded, and the funder's priorities from the RFP she uploaded last week.

Now imagine the executive director logs in before a board meeting. The assistant has a summary of this month's fundraising activity, flagged three lapsed major donors who haven't given in 14 months, and drafted talking points based on the strategic plan it's been tracking all year.

Three people. One system. Shared memory. No one typed a clever prompt. The AI just knew the work because it had been doing the work alongside them.

That's the difference between the 92% and the 7%.

The 92% are asking AI questions. The 7% built AI into their operations.

And here's what surprises people most: this doesn't require enterprise software or a six-figure budget. It requires intentional setup with the tools most nonprofits already pay for.

The prompting era served its purpose. It got people comfortable. But comfort isn't capacity.

I have a confession. I almost did not write this week's newsletter.

Not because nothing happened. Because too much happened. And by Thursday night I was sitting at my desk staring at a blank screen trying to turn the week into a lesson.

But not everything has to be a lesson.

Sometimes a good week is just a good week.

So instead of a long post, here is what I am taking into the weekend:

The weather is getting better. I took some nice walks outside and actually went to the gym after recovering from a bone bruise. A stranger replied to something I wrote online and said it helped them think about their work differently. I made dinner on a night I normally would have ordered in. And I had a conversation with someone that reminded me why I started doing this in the first place.

That is it. No frameworks. No takeaways. No "here is what nonprofits can learn from a walk in sub zero weather."

Just a good week.

I hope yours was too. And if it was not, next week is right there.

Happy Friday.

In Case You Missed It

The Heaviest Weight in the Gym. Nobody loads 300 pounds on day one. But I watch nonprofits do the equivalent with AI every week. A CrossFit lesson about progressive overload, and why the organizations making real progress started with the one task they could actually lift.

The More I Use AI, the More I Want to Talk to a Human

That sounds like a contradiction. It's not.

I use AI every single day. For research, for writing, for analysis. It's made me better at my work. I'm definitely not a skeptic.

But here's what I've noticed. The places where AI frustrates me most are the places where I used to be able to talk to a person who could bend the rules a little.

Think about your phone company. You call, you explain your situation, and sometimes the person on the other end says "you know what, let me see what I can do." They waive the fee. They make an exception. Not because the system told them to. Because they heard something in your voice.

Now imagine that's AI. You get the policy. You get the script. You get efficiency. What you don't get is someone who decides to help you anyway.

Nathaniel Whittemore wrote something recently that nails this. He was stranded in Brazil with two small kids after an emergency landing. AI was incredibly useful throughout the ordeal. But every moment that actually mattered came down to a human choosing to bend a rule. He called it "the paradox of perfect compliance."

I think about this constantly in the nonprofit world.

We have some of the most passionate, mission-driven people on the planet working at nonprofits. And too many of them are stuck behind a screen. Writing reports. Updating databases. Formatting spreadsheets. Doing work that keeps the lights on but keeps them away from the people they got into this work to help.

That's backwards.

The answer isn't to replace those people with AI. It's to put AI behind the scenes so those people can get back to the work that actually matters. Answering the phone when someone calls for help. Sitting across from a donor and listening. Walking into a community and being present. Completing the mission they signed up for.

Nonprofits don't need to become more efficient. They need to become more human. AI can help with that, if we use it to free people up instead of phase them out.

The pendulum is swinging. The people who've used AI the most are the first ones saying it: what we actually want is more human connection, not less. Let's build toward that.

In Case You Missed It

Breaking the Silo: Why Nonprofit Collaboration Is a Design Problem. I shared a personal story about writing reports that disappeared into shared drives and what it taught me about why nonprofits struggle with collaboration. Spoiler: it's not a people problem. It's a design problem.

Anyone Can Fundraise AI Weekly #16. A fundraiser told me she would have walked into a donor meeting completely cold. Her AI agent built the brief in 20 minutes. That's the shift.

Last week I sat with a fundraiser who had a donor meeting in two days and no prep done.

Not because she didn't care. Because she's a team of one, running events, writing appeals, and managing a board. Research always falls to the bottom of the list.

In 20 minutes, her AI Assistant pulled together a prospect brief using tools her org already had. Background, connections, giving history, talking points. The kind of prep that usually takes half a day or just doesn't happen at all.

But here's what made it work. The agent didn't just pull data. It knew her voice. It understood the donor's history with the organization. It remembered context from previous conversations and research she'd already done. The brief didn't read like a report from a machine. It read like something she would have written herself if she'd had the time.

That's the pattern I keep seeing. It's not that nonprofits lack talent. It's that their people are stretched so thin, the important work gets skipped.

Every org has someone figuring this out quietly on their own. Tinkering with AI after hours, testing what works. But if that person doesn't have space to bring the rest of the team along, the whole organization stays stuck.

And here's the thing. Tinkering alone won't get you there. Everyone is playing around with AI right now. Asking questions, testing prompts, seeing what it can do. That's a fine place to start, but it's not where the real shift happens.

The real shift happens when you put AI assistants and agents to work on actual tasks. Real donor research. Real proposal drafts. Real campaign planning. Not a test run. The real thing.

It's like working out. You don't get strong from reading about fitness. You get strong from reps. And the organizations putting in steady reps right now will be the ones ready when the next wave hits.

When the economy turns and corporations cut jobs, nonprofits are the safety net. They show up. But they can only do that if they have access to the same tools everyone else does.

The time to build that muscle is now.

The Tools You Use Today Won't Be the Tools You Use Tomorrow

The tools your nonprofit uses today are not the tools you will be using tomorrow.

I know that is uncomfortable to hear, especially if you just spent two years and six figures migrating to a new CRM.

But here is what I am seeing.

The major platforms that have powered nonprofits for decades, Blackbaud, Salesforce, DonorPerfect, they all share one thing in common. Their biggest selling point has become "we are secure." Not "we will transform your fundraising." Not "we will save you time." Just "your data is safe with us."

And even that is up for debate. Blackbaud's data breach in 2020 affected thousands of nonprofits and millions of donor records. The one thing they were supposed to guarantee, they couldn't.

So where does that leave us?

We are in the middle of a paradigm shift. And the organizations that pour everything into complex SaaS infrastructure right now, the ones going deep into customizations, integrations, and multi-year implementation plans, they are building for a world that is already changing underneath them.

I am not saying throw out your CRM. I am saying be careful about how deep you go.

Because the tech that survives this shift is the tech that connects to AI. That is the dividing line.

Blackbaud figured this out. They partnered with Anthropic. That was not a nice-to-have. That was survival.

What we are going to see in the next few years is a move toward simple, secure holding spaces for your data. Think less like Salesforce and more like an S3 bucket or Google Cloud Storage. No bells and whistles. No fancy dashboards. Just a clean, secure place where your data lives, connected to everything.

The intelligence does not live in the storage anymore. It lives in the AI layer on top. The storage just needs to do one thing well: keep your data safe and let AI access it.

Right now, most nonprofits are sitting on enormous tech debt. Years of workarounds, unused features they are still paying for, and processes that require a specialist to maintain. That is not nimble. That is a liability.

The pattern matters more than the platform.

If you are making a technology decision right now, ask yourself: does this make us more connected to AI, or more dependent on a system that may not exist in its current form five years from now?

That is the question worth answering.

The Only Way to Actually Learn AI

Here's something I've been reluctant to admit: You can't read your way into understanding AI.

I've tried. Spent hours consuming articles, watching demos, nodding along to webinars and studying for certifications. And then I'd sit down to actually use the thing and realize I knew nothing.

The real learning? It happens when you implement.

This is why I've become such a believer in cohorts and pilots. Not because they look good on a grant report or make for nice case studies. Because they're the only thing that works.

I've watched nonprofit professionals struggle with AI tools in our cohort sessions. Genuinely struggle. The technology moves so fast it feels clunky sometimes. You give it an instruction and it does something completely different. You try again. It hallucinates a funder that doesn't exist. Frustrating doesn't begin to cover it.

And then, without warning, it does something extraordinary.

One session, we were working on prospect identification for a participant's fundraising pipeline. The AI surfaced connections and giving patterns that would have taken their team weeks to compile manually. The room went quiet. That's when you know something clicked.

Here's the thing about pilots: they show you what works for you. Not what works in theory. Not what some vendor promises in a demo. What actually works in your specific context, with your team, with your data, with your capacity.

And here's the secret sauce: working together.

When you put a group of smart people in a room (even a virtual one), the learning curve steepens in the best possible way. Someone discovers that toggling off web search in Microsoft Copilot dramatically changes the output. Someone else figures out the exact phrasing that gets the AI to stop being so generic. Small things. The kind of things you'd never find in a manual.

Little details are what make the experience better. And you only discover them through implementation.

Without pilots, without cohorts, without actually doing the work together, the status quo just continues. Another year passes. The same conversations about "maybe exploring AI" happen at the same staff meetings. Nothing moves forward.

The organizations that will thrive are the ones willing to get their hands dirty now, even when it's messy.

I've been hitting CrossFit 4+ times a week for months now. I'm stronger than I've ever been.

I also gained weight over the holidays.

Two lessons hit me:

1. Consistency isn't everything. I kept exercising through the holidays. But when my eating got extravagant, the scale moved anyway. The work matters, but so does everything around it.

2. Progress requires constant effort. Even one week off and I feel it. Strength doesn't stick around just because you earned it once.

Here's what made the difference: having trainers and staff who helped me push through plateaus, corrected my form before I hurt myself, and kept me accountable when motivation dipped. But they couldn't do the reps for me. We had to work through the hard days together.

This is exactly what I see with nonprofits adopting AI.

Organizations that start using AI tools will absolutely get better at them. But they'll also hit pitfalls. They'll build bad habits. They'll plateau. They might even go backward during busy seasons when "good enough" takes over.

The rough patches are coming. The question is whether you work through them alone or with someone beside you.

The organizations that pull ahead won't just be the ones using AI. They'll be the ones with a guide helping them navigate the setbacks, build sustainable habits, and push through when it gets hard.

You don't need to figure this out alone.

P.S. Writing on LinkedIn works the same way. Without consistent, well-planned content, you fall off the wagon. I'll be honest, I did exactly that the last couple of weeks. Getting back on it today.

What a morning.

We just wrapped "Unlock Microsoft AI for Fundraisers" with Bob Sullivan from InfoGrow, and the energy was great. Thank you to everyone who joined us live.

Quick recap of what we covered:

The 3 levels of Microsoft AI (because yes, they name everything Copilot and it gets confusing):

Level 1: Copilot Chat. Draft emails, summarize grants, analyze donor lists. Available with basic Microsoft 365.

Level 2: M365 Copilot. Connects to YOUR data. Your SharePoint, your Teams meetings, your emails. No more uploading files one by one.

Level 3: Copilot Studio. Build custom agents for specific jobs. We demoed a Grant Discovery agent and a Prospect Finder that maps warm paths to potential donors through your board and staff networks.

The big idea: AI isn't about replacing people. It's about freeing up your team to do the relationship-building work that actually moves the needle.

Recording coming soon. We're cleaning it up and will send it out within 24 hours.

Up Next: More Free Workshops

Once we thaw out, we've got more coming your way. Virtual now, in-person later.

Smarter Fundraising with AI. With Vicky Winkler (The Marketing Shop). February 5 at 11:00 AM ET or February 12 at 12:00 PM ET. Real workflows for donor outreach, appeals, campaign promotion, and reporting. Clear use cases you can adapt right away.

No hype. No theory. Just practical steps you can apply the same week.

See you in February.

The pattern matters more than the platform.

Happy New Year. 2026 is already moving fast, and I am excited to see what trends emerge and how we can build real capacity together. This is the year nonprofits stop experimenting with AI and start implementing it, so we can better serve the communities that truly need our help.

Being Platform Agnostic Actually Means Something

This weekend I built a grant discovery demo using Google Gemini and Workspace. Last month it was Microsoft Copilot. Before that, Claude. The platform does not matter. What matters is delivering full working implementations within the tools nonprofits already use, not forcing them to switch.

It is not about flashy demos. It is about real results.

The Simpler Second Brain

Everyone is talking about building a "second brain" in 2026. Most examples I see involve Zapier, Notion, a CRM, and a dozen integrations stitched together. My approach? Claude connected to a CSV. No extra fees. No maze of tools.

For nonprofits without the budget for SaaS subscriptions or time for complex automations, this matters. Start simple. Evolve as you go.

The Unlock Most People Miss

You do not need to master every AI tool. You need to understand the pattern: an AI assistant with the right instructions and access to the right context.

Build one workflow (prospect research, grant discovery, reporting) and the same approach applies everywhere. The first one takes effort. The second takes a fraction of the time. By the third, your team sees possibilities everywhere.

The Future Is Fewer Apps

A year ago I thought building software was the answer. I spent months on a dashboard. It worked. Then I discovered I could connect AI directly to HubSpot and Salesforce without building another interface.

The future is not more apps. It is fewer apps that actually talk to each other. Before you buy that new tool, ask: can I make what I already have work better?

Want to build your first AI workflow in Q1?

I am taking on 3 pilot engagements this quarter to help nonprofits build one working AI workflow, prospect research, grant discovery, or donor communications, using the tools you already have.

Happy New Year's Eve.

I'll be honest: I thought about skipping this week's newsletter. It's the last day of the year, most of you are (hopefully) not working, and I've been fairly serious lately.

But consistency matters. And there's something nice about closing out a year the same way you spent it: showing up.

So here's the short version: 2025 was the year AI stopped being theoretical for nonprofits. I got to work alongside organizations actually building the muscle, not just talking about it. That's been a privilege.

2026 is going to move faster. If you're feeling behind, you're not. You're right on time. The organizations winning aren't the ones who started first. They're the ones who started right, with practical use cases and teams that actually know how to use the tools.

More on that in January. For now, rest. You've earned it.

This Week's Posts

Who's actually preparing communities for the AI transition? The real questions about AI aren't technical. They're human.

What I'm seeing in my work with nonprofits. The organizations pulling ahead aren't buying more tools. They're making existing tools talk to each other.

To every nonprofit professional working through the holidays. I see you. That role is only going to become more important.

See you in 2026.

At the start of 2025, AI was everywhere in the conversation. By the end of the year, adoption was still uneven.

Some nonprofits made real progress. They embedded AI into everyday workflows and freed up capacity to focus on mission. Others stayed on the sidelines, unsure where to start or worried about getting it wrong.

Both responses made sense. This year showed that AI is genuinely useful and genuinely complicated at the same time.

Here is what I actually saw.

For some, AI meant more capacity for impact.

Nonprofits that leaned in did not just save time. They redirected that time toward the work that matters: deeper donor relationships, stronger programs, more people served. Prospect research that once consumed a development officer's week now took a morning, leaving the rest of the week for actual conversations with donors.

Grant writers got something even more valuable than faster drafts. With AI handling the initial structure and boilerplate, they had time to actually connect with program managers and officers before writing. They could ask better questions, understand the nuances of the work, and craft narratives grounded in real relationships rather than rushing to meet deadlines. The grants got stronger because the process got more human, not less.

For others, the trepidation was real.

Many nonprofits I spoke with were cautious. Some were overwhelmed by the options. Others had tried a tool, got inconsistent results, and stepped back. A few were waiting for clearer guidance from their boards or funders before moving forward.

This hesitation was not irrational. AI does hallucinate. It does miss context. It can create more work if you do not know where it fits and where it does not.

The gap between the organizations moving forward and those holding back was not about budget or tech savviness. It was about having a clear starting point and support to navigate the learning curve.

The biggest wins came from tools nonprofits already had.

The organizations seeing results were not chasing the newest platforms. They were using AI inside Google Workspace, Microsoft 365, ChatGPT, Anthropic, Copilot, and Gemini.

AI worked best when embedded into everyday workflows like Docs, Excel, email, CRMs, and dashboards. Less context switching. More momentum.

AI improved the work environment, not just the work output.

This is the part that does not get enough attention. Nonprofit professionals are burning out. Small teams carry impossible workloads. The sector loses talented people because the daily grind overwhelms the mission-driven purpose that brought them in.

When AI is implemented well, it does not just improve productivity metrics. It improves the lived experience of the people doing the work. Less time on repetitive tasks means less exhaustion. More capacity means less guilt about what is not getting done. Sustainable workloads mean people stay longer and bring their best.

Healthier organizations serve their communities better. That is the real return on investment.

Capacity building mattered more than tools.

2025 made one thing clear. Access to AI is not the hard part. Knowing how to use it in ways that actually build capacity is.

The biggest lesson from 2025: AI does not replace nonprofit professionals. It protects them from burnout and frees them to focus on impact, when used with intention. Used without intention, it just adds to the pile.

2026 will not be about whether nonprofits should use AI. It will be about who has the support to use it well.

The Pivot That Changed Everything

In early 2024, I spent 4 months building a platform called ACF Bridge.

It did everything I thought it should. Analyzed notes. Surfaced insights from contacts. AI-first philosophy. It worked. I was expanding capabilities daily. Campaigns, templates, dashboards.

Then MCP servers were born.

Model Context Protocol. Basically, connectors that let AI talk directly to your existing tools.

I looked at everything I'd built and realized: I can do all of this without building another interface.

I could analyze my email, my contacts, my CRM. Without asking anyone to learn a new platform.

That was the hardest moment of my year. Four months of work. Functioning software. And I had to let it go.

But here's what I learned: The goal was never the platform. The goal was helping nonprofits work smarter.

So I iterated. Built a workflow system that produces the same results but grows with technology instead of fighting it. No technical debt. No heavy infrastructure. Just flexible tools that adapt.

The lesson: Don't fall in love with what you built. Fall in love with the problem you're solving.

This week's posts all circled the same idea: knowing when to hold on and when to let go.

I've joined the 3am club this year. Every weekday I wake at 3am. Mind racing. 72% of entrepreneurs deal with mental health challenges. Building something new is not simple. But the mission is worth it.

The Holiday Email Paradox. Every year since 2018, I write a personal holiday email. Hand-written. AI can help you find 20 prospects in 30 minutes, but it can't wrap your memories in words the way you can. Some things should stay slow.

The best system I ever built was 5 HTML dashboards. The best system I'm actually using? Two CSV files. The fanciest tool you never open loses to the simple one you use every day.

The pattern underneath all of this: This week was about letting go. Letting go of the platform I built. Letting go of the "perfect" system. Letting go of the idea that AI should automate everything.

The answer is almost always: start simpler than you think. Build for the problem, not the solution. And know when something isn't working, even if you love it.

Real AI. Real Work. Real Humans.

This week's theme was simple: AI helps, but humans still do the work that matters.

AI doesn't close grants. Humans do. AI isn't a silver bullet. It doesn't replace strategy, relationships, or judgment. But it can remove friction so fundraisers can focus on the work that actually raises money. If AI isn't freeing up time for human work, you're using it wrong.

Certification: AI Fluency for Nonprofits. Completed Anthropic's AI Fluency program to validate ACF's approach to responsible AI adoption. The big lesson? The best systems keep humans firmly in the loop. AI should augment judgment, not replace it.

Email is dead. Phones still work. Outreach math from the field: 100+ emails = 3 replies. 20 phone calls = 8 conversations. AI may scale communication, but trust is still built one conversation at a time. Automation doesn't build relationships. People do.

Your Development Director is drowning in December. December is survival mode for nonprofit teams. Appeals, reports, research, gratitude, planning. All at once. The solution isn't "work harder." It's removing one major time-drain with automation. Pick ONE task. Automate the first draft. Give your team air.

The pattern underneath all of this:

AI is not here to: replace fundraisers, remove judgment, or eliminate expertise.

AI is here to: remove busywork, eliminate friction, accelerate preparation, and protect human energy.

When implemented correctly, AI doesn't make nonprofits colder. It makes them more human.

This week's theme: AI works best when it follows humans, not when it pretends to replace them.

This week across Anyone Can Fundraise, one idea showed up again and again: AI doesn't generate impact. People do.

Context beats cleverness. I explored the difference between novelty AI and practical AI. Snark is entertaining. Context is valuable. Real results happen when AI understands your work, your systems, and your world, not just your words.

You cannot replicate lived experience. AI cannot replicate you. Without human input, it doesn't grow, adapt, or improve. It stays exactly as it was. People evolve. AI doesn't. That's why the future isn't replacement. It's collaboration.

Personal brand is usefulness, not polish. Brand isn't performance. It's contribution. If you help people move faster, think clearer, or feel less stuck, that's your brand.

You already own the tools. Most nonprofits already have AI in their stack. The real challenge is not procurement. It's enablement. Progress comes from use, not upgrades.

The takeaway this week: AI doesn't win by being smarter. It wins by being embedded. In your systems. In your decisions. In your real work.

That's the kind of AI we build at Anyone Can Fundraise. Human-led. System-aware. Actually useful.

Welcome back to the Anyone Can Fundraise weekly roundup. This week was a big one. Not because of a single moment, but because the speed of everything has changed.

I Built a Nonprofit Game in Minutes Using AI. I took a nonprofit idea, talked it through out loud, turned that conversation into a structured canvas, and then dropped it into Google AI Studio using the new Gemini 3.0 model. In one prompt, the full game was built. The Nonprofit Quest puts you in the shoes of a small fundraising team starting with limited time, limited staff, and limited resources.

From ChatGPT as a Curious Friend to a Fundraising Copilot. In 2022, ChatGPT was a novelty. Today, it is one part of a rapidly expanding ecosystem. This week I shared a story about a nonprofit director who implemented the Prospect-Finder workflow I built for her using Microsoft 365. We never even got on a call. Six days later she messaged me: "I just got your Prospect-Finder tool working. Thank you!" She went from searching for donors manually to identifying warm prospects in minutes.

Cursor as a Business Command Center. Another experiment from this week: using Cursor not as a coding IDE, but as a business operations hub. I used it to organize my entire file system, clean and consolidate folders, build structured dashboards inside Google Drive, spin up content and research agents, and create order from scattered digital chaos.

The Toughest Feedback of the Week (From an AI). One post that sparked strong conversation was my story about asking Claude to identify my blind spots. I expected minor notes. Instead, it delivered a straight shot of honesty: "You are over-engineering proposals. Clients do not need complex frameworks. They need clear outcomes."

How These Threads Connect: Capacity unlocks when you combine AI with clear workflows. Whether it is turning a spoken idea into a working game, helping a fundraiser find prospects in minutes, building a digital command center, or simplifying client work to essentials, the pattern is the same. The tools are here. Most nonprofits already have access to them. The real power comes from knowing how to set them up the right way.

That is the mission behind Anyone Can Fundraise.

If there's one thing that's becoming clear, it's this: nonprofits aren't just part of the solution, they are the solution to many of the challenges shaping our future.

As businesses restructure and governments pull back, nonprofits are stepping in where the gaps are widest. They're not just providing services, they're rebuilding trust, resilience, and local capacity in a time when both technology and society are changing faster than anyone expected.

This week, my posts looked at what that means in practice from practical AI tools to the human side of progress and how we can all help these organizations build the capacity they need to keep going.

Highlights from this Week

Progress, Not Perfection. Change doesn't happen through perfection, it happens through persistence. This post explored how small, steady progress in adopting AI and better systems can make a bigger impact than sweeping overhauls.

Google Labs, AI for Good and Nonprofit Innovation. Google's AI experiments show how accessible tools can empower nonprofits to test ideas safely, creatively, and at low cost, proving innovation doesn't need to start with a massive grant.

Why Nonprofits May Be Part of the Answer. In a world where job loss and automation dominate the headlines, nonprofits are a stabilizing force, giving people purpose, structure, and connection when it's needed most.

AI for Good: Practical Tools for Fundraisers. AI isn't the enemy of fundraising, it's the efficiency multiplier. This week, I shared a first look at a marketing and outreach tool I've been developing to help fundraisers automate their day-to-day communication.

The Big Picture: If we want a strong future, we can't treat nonprofits as a safety net. They're part of the foundation. The imperative is clear: we need to invest in their capacity, their people, and their technology, not someday, but now.

And that starts with testing, sharing, and building together. Whether it's new fundraising tools, AI workflows, or capacity-building playbooks, your feedback helps shape what's next.