Promethean for Good · mission tier
AI in humanitarian work
deserves the strongest audit layer.
Refugee triage. Crisis response routing. Medical triage in mass-casualty events. Climate-emergency early warning. When AI touches vulnerable populations, the audit trail isn't a nice- to-have. It's the contract.
Humanitarian organisations are increasingly deploying AI in decisions that affect people in vulnerable situations. Triage at a refugee reception centre. Resource allocation in a famine-response operation. Medical decisions when hospitals are overwhelmed. Climate-disaster early-warning routing. The technology works. The audit layer is missing.
The same regulatory frameworks the substrate addresses for commercial operators — GDPR Article 22, the EU AI Act, MDR for medical decision support — apply to humanitarian work too. Often more sharply: the people affected have less recourse if something goes wrong. The audit layer is more important here, not less.
Promethean for Good is structurally €0. Not a discount, not a pilot grant, not a sponsorship. The substrate's marginal cost per humanitarian deployment is near zero; the compounding returns from these deployments — trust, reputation, recruitment — are real but indirect, and we don't bill them back. Who qualifies →
Why a mission tier exists
Mission alignment is the cheapest cost-of-capital we have. Saying so plainly.
What humanitarian orgs get
- The full substrate — same Phase R / v1.2 architecture our commercial operators use
- L12 receipt chains anchored to Bitcoin via OpenTimestamps, signed under Promethean's trust key
- Substrate-engineering support (40 hrs / quarter): co-design of RuntimeAISpecs, reviewer-gate predicates, production deployment, regulator-readable handoff
- Joint engagement with humanitarian-sector regulators (EU AI Office, national DPAs, MDR notified bodies, sectoral inspectors) when context warrants
- Reference-deployment status: humanitarian deployments demonstrate the substrate at its highest-stakes
What Promethean gets (stated plainly)
- Reference cases that operators and regulators both respect — humanitarian deployments prove the substrate at the limit, not in the lab
- Recruitment magnet: the engineers we most want to hire want to work on AI that touches vulnerable populations, done right
- Regulatory goodwill: the EU AI Office cares about humanitarian AI; serving it well at zero cost is a credibility signal
- Sharper architecture: humanitarian workloads stress-test the substrate in ways commercial pilots never will (intermittent connectivity, multi-jurisdiction tenancy, extreme reviewer-queue latency tolerance, etc.)
- Mission alignment: this is not a side-effect of the architecture. The substrate exists because AI deserves better audit — that applies everywhere, especially here
Who qualifies
Four criteria. All four required.
Verified non-profit status
Registered charity (UK, EU member state, US 501(c)(3), or equivalent jurisdiction). UN agencies and inter-governmental humanitarian bodies qualify.
AI touches vulnerable populations
The AI feature you'd deploy through the substrate is directly involved in decisions about refugees, asylum seekers, disaster victims, people in medical crisis, citizens in emergency situations, or other vulnerable groups.
Mission-fit interview
One conversation with substrate engineering. We confirm the deployment matches the spirit of the tier (not just letter-of-eligibility), and we map the regulatory frameworks that apply to your specific context.
Capacity available
We cap at five active Promethean-for-Good deployments at any time. If you'd be the sixth, we hold your application and re-open intake when capacity frees. We never compromise on the engineering depth each deployment gets.
Use cases we're built for
Four sectors. Each carries serious downside risk if AI gets it wrong.
Refugee + asylum operations
Examples
UNHCR-style reception triage · multi-language intake routing · resource-allocation prioritisation · family-reunification matching
Regulatory context
GDPR Art. 22 (automated decisions affecting individuals) · EU AI Act Annex III (migration / asylum decisions are explicitly high-risk)
Crisis + disaster response
Examples
Red Cross-style needs assessment · damage classification from imagery · supply-chain coordination · volunteer dispatch routing
Regulatory context
National emergency-response frameworks · sectoral inspector oversight · post-incident review by independent commissions
Humanitarian medical
Examples
MSF-style triage in mass-casualty incidents · supply-chain forecasting in low-resource clinics · medical-image pre-screening in conflict zones
Regulatory context
MDR (Medical Device Regulation) class-IIa+ for decision-support · AI Act Article 6 (high-risk AI in healthcare) · clinical-ethics oversight
Climate + ecological emergency
Examples
Early-warning routing for extreme-weather events · evacuation prioritisation · damage-prediction modelling · environmental-violation detection
Regulatory context
NIS 2 (critical infrastructure including emergency services) · sectoral environmental regulators · public-accountability journalism
Boundaries, plainly
What Promethean for Good is not.
Not
Not a sponsorship
This is full substrate access at zero cost. There's no co-branding requirement, no PR obligation, no demand for case-study rights beyond the reference-deployment line item.
Not
Not a startup-discount knockoff
We don't offer 50%-off for non-profits. The price is €0 because the alternative (charging a stripped-down version) is worse for everyone involved.
Not
Not a way to deploy AI you shouldn't
The substrate makes AI decisions auditable. It doesn't make a bad-decision AI feature good. We won't help deploy something we wouldn't deploy ourselves.
Not
Not an exit ramp from compliance
GDPR, AI Act, MDR still apply to humanitarian deployments. We help you satisfy them; we don't pretend the substrate substitutes for legal review.
Not
Not a perpetual commitment
If a humanitarian deployment converts to a commercial product (e.g., a charity spins out a for-profit), the engagement moves to commercial pricing. We're transparent about this from intake.
Not
Not a marketing channel
We don't promote Promethean for Good deployments in our marketing without the org's explicit, ongoing consent. Reference-deployment status is opt-in per use case.
Apply
One email. Four things in it.
Send to · for-good@promethean.software
- 01
Your organisation
Name, legal status (registered charity / 501(c)(3) / UN agency / etc.), and a one-paragraph mission statement.
- 02
The AI feature you'd deploy
What it does · who it affects · what decision it informs · what currently happens without it. One page is plenty.
- 03
The regulatory context
Which frameworks apply to your deployment (GDPR, AI Act, MDR, NIS 2, sectoral). Your guess is fine — we'll map it properly together in the mission-fit interview.
- 04
A reference contact
One person in your organisation with authority to commit (or escalate). We don't need procurement involvement at the application stage — that comes later if at all.
Response time: within 7 days. If capacity is currently full we tell you immediately and confirm when intake re-opens. No "we'll get back to you" voids.
“Humanitarian AI without an audit layer repeats the worst failure modes of humanitarian work at software speed. The substrate exists so that doesn't have to happen.”
Apply
for-good@promethean.software
One email, four things. 7-day response. No procurement maze.
Other tiers
See full pricing
Four self-service tiers from Dev (€0) to Enterprise (€2,499/mo), plus partnership tiers. The For Good tier sits alongside, not below.
The substrate
What you'd actually get
The Phase R / runtime-AI containment one-pager. Same architecture commercial operators use. Seven reliability pillars, 243 tests.