Every consultant knows the pro bono math.
You have skills. Organizations need them. You want to help. But every hour you donate is an hour you don’t bill. So you limit your generosity to what your revenue can absorb, feel vaguely guilty about not doing more, and move on.
I’ve done that calculation for fifteen years. Last weekend, the math broke.
I built a full community events platform for a pro bono client — strategy, development, security hardening, deployment to production — in one weekend. Not a prototype. A real application at whatcomconnects.org that aggregates civic and community events across Whatcom County, with 13 local organizations seeded, a public event submission system, and a one-click admin approval workflow designed for a non-technical volunteer.
Total investment: $7.68 for the domain name. Everything else runs on free tiers.
And that changes the math on everything.
The Project That Wasn’t Supposed to Be This Project
Here’s the consulting reality that AI can’t replicate: the client didn’t ask for a community calendar.
A friend connected me with a local civic organization that needed help with graphic design, social media, and their website — the classic volunteer org capacity gaps. Discovery revealed someone else was already handling their website. But the conversation surfaced a bigger problem: Bellingham’s progressive community has dozens of active groups and no central place to find events. People were checking six different Facebook pages, a partially-maintained website, and word of mouth.
The ask was social media help. The actual need was civic infrastructure.
So I pivoted. Instead of patching the original organization’s communications, I proposed a standalone, neutral community calendar — separate branding, separate positioning, open to any local organization. That pivot required zero technical skill and fifteen years of knowing that the first ask is rarely the real problem.
Then I built it. In a weekend.
Strategy First, Code Last
Here’s what I didn’t do: open a terminal and start typing.
Here’s what I did: I spent hours writing structured specification documents. A product spec that mapped user stories, data models, and submission workflows. A visual design brief specifying the PNW-inspired color palette and typography. A security requirements doc that addressed a constraint most developers wouldn’t think about: the admin volunteer’s identity needed to be completely shielded from the public-facing system because of the political context in the community.
Then I handed those specs to Claude Code — and let it build.
This is the pattern I keep discovering: the more specific you are up front, the faster it goes. If I’d typed “build me a community calendar,” I’d have gotten something generic requiring days of revision. Instead, I handed Claude Code a thoughtful brief and got a thoughtful product. Thirteen organizations. Sixteen real events scraped from their websites. An approval workflow where the admin volunteer receives an email, taps Approve or Reject, and she’s done — zero training required.
I keep finding this in every project: the bottleneck is never the code. It’s knowing what to build. The PwC research I wrote about last week puts a number on it: process redesign delivers 80% of AI’s value. Technology delivers 20%. That ratio played out precisely.
What AI Built vs. What I Caught
I want to be honest about something: Claude Code made mistakes.
On the first pass, it set overly permissive database access policies — the kind of shortcut that works in a demo but creates real security vulnerabilities in production. It marked database functions with the wrong volatility settings, which would have returned stale data. It put a personal email address in documentation because it appeared in the project context.
I caught all of this because I knew what to look for. Not because I’m a security expert — because I’ve spent years managing projects where “it works” and “it’s safe to deploy” are very different standards.
This is the part of AI-assisted building that doesn’t make the hype cycle: the human review pass isn’t optional, and the quality of that review depends entirely on experience. A junior developer would have shipped those permissive database policies. Someone who’s managed production systems and client data knows to check.
The workflow that actually works isn’t “AI builds it.” It’s: strategic decisions → structured spec → AI executes → human reviews for things AI doesn’t know to worry about → targeted fixes → ship. The AI was indispensable. The judgment was irreplaceable.
The Real Economics of Generosity
Traditional pro bono consulting operates on scarcity math. Compass Pro Bono, one of the largest matching organizations in the US, has facilitated $120 million in donated consulting services since 2001 — through carefully managed 4-8 month engagements with hand-picked teams. Deloitte professionals donate over 158,000 hours annually. These are impressive numbers built on a fundamental constraint: skilled time is expensive, so you ration it.
But what happens when the build time drops 90%?
The old math: A community platform = 3-6 weeks of development = $15,000-$40,000 in opportunity cost. Something you’d need a grant or significant budget to justify.
The new math: A community platform = one weekend + $7.68 for the domain. Zero monthly cost. Zero ongoing operational commitment.
When building costs a weekend instead of a quarter, you don’t need a committee or a formal engagement or a 4-month timeline. You can just help. The economic barrier that made generosity a sacrifice doesn’t disappear — it becomes so small that the old excuses stop working.
What This Means for Each of You
If you’re running a small business: You’ve said no to building something for your industry association, your chamber, your neighborhood group — because you couldn’t justify the time. What if the time cost was a weekend? What community tool have you been sitting on because the build felt too expensive? With free-tier infrastructure (Vercel, Supabase, Cloudflare), you’re looking at single-digit annual costs, not monthly invoices.
If you’re a marketing leader: Your team has ideas for internal tools and community resources that keep getting deprioritized because “we don’t have dev resources.” The strategy-first approach means your team’s business knowledge is the development resource. The people who understand the problem, the user, and the constraints are more valuable than the people who write the code — because AI writes the code now.
If you’re a solo entrepreneur in the Netherlands (or anywhere you’re building community from scratch): Pro bono work is one of the best ways to build professional credibility and local roots. But when you’re a one-person operation, every donated weekend is a weekend you’re not building the business that keeps your visa valid. When a complete, deployed project fits in a single weekend with $7.68 in costs, that tradeoff becomes almost trivial. One weekend of community contribution. Months of relationship capital.
The Uncomfortable Implication
Here’s the part that might sting: if you can build meaningful things in a weekend now, and you’re not using any of that reclaimed capacity for something bigger than your own revenue — what exactly are you saving it for? I asked a version of this question five months ago about automation — what happens after you’ve saved all those hours? This is the answer I didn’t have then.
The latest Small Business Expo survey (n=693, published last week) found 78.6% of small businesses report measurable AI efficiency gains. Business.com’s 2026 research puts the average at 5.6 hours saved per week. That’s not theoretical. That’s time showing up on your calendar right now.
The question isn’t whether AI saves you time. It’s whether you’ll spend any of that time on something that doesn’t directly generate revenue — and whether that might, paradoxically, be the most valuable thing you do all year.
What I Actually Learned
The tactical lessons were predictable: Claude Code is fast, Vercel deployment is painless, and structured specs produce dramatically better output than vague prompts.
The strategic lesson was not. The most underrated thing AI changes isn’t productivity. It’s the economics of generosity. When building approaches zero-cost, the constraint shifts from “can I afford to help?” to “do I know what would actually be useful?” And that question rewards business experience, community knowledge, and judgment — exactly the things AI can’t provide.
The best thing I’ve built with AI wasn’t for a client or for my business. It was a community calendar for civic organizations that couldn’t afford a developer.
Total cost: $7.68, one weekend, and fifteen years of knowing that the first ask is rarely the real problem.
Where This Gets Practical
WhatcomConnects exists because someone mentioned they wished local events weren’t scattered across six different Facebook pages. Most organizations I work with have a version of this — a tool, a platform, a resource they know would help but can’t justify building.
If you’re sitting on one of those ideas and aren’t sure what it would take to make it real, that’s the kind of conversation I have. No pitches, just a realistic assessment of what’s buildable now that wasn’t six months ago.
Want more like this? I write about AI adoption for people who run things — not hype, not doom, just what’s actually working. Subscribe here.
Sources:
- Deloitte: Pro Bono and Skills-Based Volunteering
- Anthropic Research: “How AI Assistance Impacts the Formation of Coding Skills”
- SemiAnalysis: “Claude Code is the Inflection Point” (February 2026)
- Small Business Expo: “Accelerating AI Adoption and Measurable Gains in 2026” (February 2026)
- Business.com: “2026 Small Business AI Outlook Report“
- Compass Pro Bono: $120M in pro bono consulting since 2001
