System Change AI
q2 ’26 · open intake talk to us
Not-for-profit product studio · est. 2026 · operating in public
v0.4 / public distributed · decentralized

Tools for the
people changing
systems.

System Change AI is a not-for-profit product studio building open-source AI infrastructure beneath the long work of shifting systems — from the first convening to the final wire transfer.

Bring us your dream See what's in the studio
Underway 04 in development
Partners 11 funders & conveners
Ship cadence 90d first useful version
Studio 2026 → operating in public
01 — why now

The systems we built are producing the polycrisis. They need to produce regeneration.

Climate, inequality, fragmentation, ecological collapse — not separate problems. They are the output of how we wired the world.

The systems running our economies, our institutions and our infrastructure weren't designed for the world they're now operating in. The damage isn't a bug — it's the consequence of incentives, structures and feedback loops we built decades ago.

The people doing the long work of shifting those systems toward regeneration deserve tools shaped to the actual shape of that work.

i. The world is in crisis. premise
ii. Our systems need to change. consequence
iii. By building the tools to support systemic changemakers, we increase humanity's capacity to be anti-fragile. our part
02 — thesis

AI is rushing toward the largest market. We're aiming it at the highest leverage.

Systems change is slow, plural & relational. That doesn't make it un-automatable — it makes the automations matter more.

The people doing the hardest work in our economy — conveners running multi-year coalitions, foundations underwriting unproven theories of change, impact funds stitching together blended capital stacks — still operate out of inboxes, decks and spreadsheets.

We don't think the answer is another dashboard. It's a generation of small, opinionated instruments shaped to the actual shapes of this work — tools that understand stakeholders, capital, narrative and time horizon as first-class objects.

So we build them. As a studio. Releasing four to six products a year, each co-owned with a partner who lives the problem.

03 — the spectrum

From the first convening to the final wire. Discrete tools. Open seams.

Systems-change work isn't a funnel — but it does have texture. Each stage produces artifacts the next stage struggles to read. We build at the seams.

04 — in studio

Four products underway. More in sketches.

P/01 · MYCELIAL
Network

The network, woven.

Mycelial maps the threads of relationship across a field — surfacing latent connections, naming what's holding a coalition together, and brokering the introductions the network didn't know it needed.

In dev · 4 networks mycelial.systemchange.ai →
P/02 · MYCREW
Sovereign agent

Your sovereign agent, on your machine.

MyCrew is a personal AI that runs locally on your laptop — sovereign, private — and quietly negotiates with the agents of your peers across organizations to catalyze the collaboration that would otherwise never happen.

In dev · alpha cohort mycrew.systemchange.ai →
P/03 · SYSTEM INTELLIGENCE
Intelligence

The system, listening to itself.

A continuously-running intelligence agent that watches activity across a field — new actors, new capital, new narratives — and keeps everyone working on the issue up-to-date on what just changed and why it matters.

In dev · 2 fields signal.systemchange.ai →
P/04 · ——
In stealth

Something, in stealth.

A fourth product is in early development with an anchor partner. We'll talk about it when it's ready to be useful. If you have a guess about what's missing, you're probably right — let us know.

Sketch · 2026 —
05 — how we work

We're a studio, not a vendor.

01 Sit in the work. Every product starts with a six-week residency inside a partner organization. We don't write a line of code until we've sat through the meetings, read the inbox, and been confused. Weeks 01–06
02 Build the smallest useful thing. A 90-day shippable v1 — narrow, opinionated, embarrassing to demo. We'd rather earn a renewal than launch a website. Weeks 07–18
03 Ship open, by default. Every tool ships under permissive open-source licenses, speaking open protocols. Anchor partners help shape what gets built; the commons gets to keep it. Ongoing
04 Hand it to the commons. When a tool needs market gravity to keep running, we spin it out as a self-sustaining entity — community-owned, governed by the people who depend on it. The studio stays in service. The tool stays sovereign to its users. Year 2+
06 — what we believe

Six commitments. Held loudly.

  1. I.Built for impact, not revenue. We're not-for-profit by design — and when something needs market gravity to live, we spin it out as a self-sustaining entity owned by the people using it.
  2. II.Your synthesized intelligence is yours. The AI that comes to know you, your work and your relationships runs on your machine, and answers only to you.
  3. III.Open source. Open protocols. Everything we build ships open and speaks open standards. We don't believe in lock-in.
  4. IV.Structurally non-enshittifiable. The incentives are wired so we cannot extract from the people we serve. No wisdom mining. No bait-and-switch. We design the failure modes out.
  5. V.Success is twenty interoperable things, not one platform — small, sharp tools that talk to each other and belong to the communities that use them.
  6. VI.The point isn't to scale the studio. The point is to retire it — by building tools the field eventually owns.