How we build

You no longer have to choose between innovation and upholding your values.

The Progress Lab delivers guidance from AI experts and provides a mission-aligned AI platform that enables progressive groups to harness the power of general purpose foundation models within a robust security and transparency framework.

Progress Platform · New task

Task: Build the monthly donor report

Modelsorted by impact

Claude Haiku 4.5	low impact	$
Open Weight 150B	low	$
Claude Sonnet 4.6	moderate	$$
GPT-5	moderate	$$
Claude Opus 4.6	higher	$$$

Energyus-west-2 · Pacific Northwest grid

Run now406 gCO₂/kWh

Clean window35 gCO₂/kWh · 7:00 AM

now+6h+12h

Scheduled for 7:00 AM ET · −91% carbon

The scheduler runs queued tasks in the grid's cleanest window — every run logged with its model and carbon intensity. Figures illustrative; the product uses live grid forecasts.

Our values

Five commitments, built into the product.

Each one is a promise — and a set of product features that make it real. Open any value to see how.

AI amplifies human judgment; it doesn't replace it.

Customizable plugins ensure the platform adapts to how your team already works
Open-ended model access to build something new without sending your prompts and data to Big Tech
Human review is built into our workflows — nothing ships without human oversight
Full audit trail — see who did what, on which model, so people stay accountable and in control

Bias, impact, and methodology live in the open.

Bias scores at the point of model selection, drawn from peer-reviewed research
Environmental impact made visible — carbon and water usage is visible
You select the model size, time-of-day scheduling, data center region to control environmental impact
We publish methodology — sources cited, scoring visible in-app
Citation-based fact-checking means every claim is traceable to a source

When security and convenience conflict, safety wins every time.

Perimeter defense — 2-factor authentication, password & access rules, encryption in transit and at rest
Access governance — role-based access management, policy enforcement, and full audit logs
Defaults, not upsells — security at the architecture level, not as configuration

Your data stays yours.

Your data, queries and prompts are never sent to third-party AI vendors
Customer-managed encryption available — you hold the keys, not us
Your data is never used to train models for Big Tech (or us!)

We're model-agnostic on purpose. The right tool serves the organization's mission, not a vendor relationship.

Choose the right model for each task, the moment, and the mission
Bias and impact surfaced so the choice serves your mission, not a vendor
No lock-in — no single-vendor dependency baked into your stack

Security · Internal & external

Safety wins every time.

Security at the architecture level, as a default — not an upsell. Three layers, from the login screen to the model itself.

The Front Door

Perimeter defense

Two-factor authentication on every account
Password & access rules enforced platform-wide
Encryption in transit and at rest

Inside

Access governance

Role-based access management
Policy enforcement built into every workflow
Full audit logs — who did what, on which model

Vendor Access

Private by design

No external AI APIs — private inference only
Customer-managed keys
Your data never trains anyone’s model