We tend the conditions.
The field does the rest.

How This Began

It started with a question

The Pollination Project has spent years making small grants to grassroots changemakers — ordinary people building kindness, care, and community in their own corners of the world. One day, they approached ServiceSpace with a question: could your technology help us build a relational field across about 6,800 grantees in over a hundred countries?

In trying to answer it, and holding its potential, the obvious came into view. It wasn’t only The Pollination Project’s grantees who wanted this. Mission-driven organizations everywhere are asking a version of the same thing — how do we turn an audience into a community, a network into a kinship? And AI, arriving fast, could carry that work into places it could never reach before — if it were pointed in the right direction. Not to accelerate the status quo, but to accelerate toward something better.

Shunya Labs is what grew from that question. A small lab, incubated by The Pollination Project, with a single brief: take hard-won design wisdom and build the AI tooling that can carry it — far past where any single movement could reach — to the public, nonprofit, and civic organizations now asking for it.

The Proof · ServiceSpace

A working model, not a theory

We did not invent the ideas underneath this work. We inherited them. For more than twenty-five years, ServiceSpace ran a living experiment under deliberate constraints — volunteer-run, gift-based, no fundraising, no paid staff, no impact metrics — to learn what designs emerge when you assume people want to give. By every conventional measure the constraints looked limiting. Held long enough, they revealed a different kind of abundance.

What they were practicing has a name: social permaculture — tending the conditions through which human goodness self-catalyzes. Visitors to a farm ask what should we grow here? The farmer asks what grows here? Masanobu Fukuoka spent fifty years on that second question — no tillage, no fertilizer, no pesticide — and his yields stunned the agronomists. The insight was precision, not laziness: one gram of his soil held a hundred million nitrogen-fixing bacteria. He didn’t put them there. He stopped killing them.

The same precision is why ServiceSpace refused to measure. When a measure becomes a target it stops being a good measure — build a dashboard for flourishing and you get a performance of flourishing, while the real thing slips out the back. Drop the metrics and you can tend the immeasurables — trust, meaning, belonging — without collapsing them into a number. That is the ground we build on:

What few people know is that ServiceSpace was always a technology story. Almost no one sees the machinery underneath — millions of lines of code, written long before the current boom, that let strangers micro-volunteer before there was Wikipedia; infrastructure that has sent two emails a second for decades and never once carried an ad or asked for anything; millions arriving each month to find nothing sold to them, no funnel narrowing toward a purchase. The machine was there the whole time, and almost no one felt it. That was the point: build the tool, then let it disappear; gather people, then let the room become the technology; trust the field before scaling the form. That instinct is the one we carry forward.

The Clock

AI is arriving fast — and pointed the wrong way

The benefit is not being shared. Roughly one in six people on earth has touched these tools at all; the number who use them fluently is far smaller, and clustered among those who already had the most of everything else. Built by the market, AI is shaped to capture attention and extract value — the receivers handed out fastest to those who already owned three more.

And it is moving so fast that the organizations who most need it often don’t even know what to ask. The window in which they can build capacity — the infrastructure, the trust, the fluency — is open now, and will not stay open forever.

So this is the work: most AI help makes you better at the system you’re already in. We help organizations make the leap instead — building the tooling on a model already proven to work, so what you adopt today leans toward the world you actually want rather than locking you deeper into the one you don’t. We want to equip organizations and movements to solve tomorrow’s challenges, not just manage yesterday’s problems.

A Worked Example · The Circle Agent

AI that composes a room — then leaves it

In a community of fifty, a host intuitively knows who should meet whom. In a community of thousands, that relational context explodes beyond any human’s grasp. The Circle Agent reads who connects with whom and what each person reflects on, in their own words, and seeds small circles of near-strangers who share an unnamed question — one community held 32,275 pairs with more than twenty shared concepts and zero shared history. It then handles the dozens of invisible coordinations a single circle needs, and fades as the humans take root.

It is one of several things we build — alongside wisdom companions, positive-media infrastructure, and the hands-on “AI Jump” that helps a nonprofit make this leap without losing its heart.

AI as the invisible mycelial network for relational transformation. It creates the conditions — and then disappears.