Project Liberty Alliance

The Network That Needs Its Own Infrastructure

The Project Liberty Alliance connects 175+ organizations working on responsible technology, digital rights, and civil society. Frank McCourt's coalition is one of the most concentrated networks of mission-driven organizations anywhere — and the team managing it runs on Airtable, a listserv, and manual coordination.

Lara Galinsky and the Alliance team are doing the relationship work that makes warm introductions, pop-up convenings, and cross-org collaboration possible. But the infrastructure underneath that work hasn't kept up with the network's growth. What started as a manageable directory is now 175+ organizations with overlapping missions, complementary datasets, and no connective tissue between them.

This proposal covers two workstreams: the internal systems that the Alliance team needs to operate, and the network-wide data intelligence layer that turns 175+ isolated organizations into something greater than the sum of their parts.

Proposal

---

Workstream 1: Internal Digital Transformation

The Problem Today

The Alliance team manages relationships with 175+ organizations using disconnected tools. Member data lives in Airtable. Communications go through a listserv. Event coordination — monthly virtual pop-ups, in-person convenings — is tracked separately. HubSpot handles some CRM functions, but it's not wired to the member directory or the engagement history.

The result: the team knows their network through personal relationships, not through data. There's no unified view of which organizations are engaging, which have overlapping needs, which should be introduced to each other, or which are drifting away. Every warm introduction requires Lara or Harrison to hold the full context in their heads.

For a team this small managing a network this large, that's unsustainable as the Alliance grows.

What We'd Build

Phase 1 — Systems Audit & Data Consolidation

Map every tool the Alliance team touches. Airtable, HubSpot, email, event platforms, the member directory. Identify what data lives where, what's duplicated, what's missing. Consolidate into a unified partner data model.

Phase 2 — Integration Middleware

Wire the systems together. When a new org joins the Alliance, their profile propagates everywhere — CRM, directory, event invites, listserv. When a pop-up happens, attendance and engagement data flows back into the member record automatically. No more manual syncing.

Phase 3 — Relationship Intelligence

Build the layer that surfaces what the team already knows intuitively but can't scale: which organizations should be introduced to each other, which are working on complementary problems, which haven't engaged in months and need a touchpoint. Light ML on engagement patterns and org profiles to augment — not replace — the team's relational knowledge.

---

Workstream 2: Network Data Intelligence Service

The Bigger Opportunity

175+ organizations in the Alliance are each sitting on data — program outcomes, research findings, operational metrics, community indicators. That data is siloed. Each org uses it for their own reporting, their own grant applications, their own internal dashboards. Nobody sees the full picture.

The Alliance is uniquely positioned to change that. Not by collecting everyone's data into a central warehouse — that raises every governance concern in the book — but by building a data intelligence layer that lets partner organizations discover what data exists across the network, understand where it overlaps and complements, and collaborate on shared datasets where it makes sense.

This is where the Alliance's relationship capital meets technical infrastructure.

What This Looks Like

Partner Data Catalog

A queryable index of what data each Alliance partner holds, at the metadata level. Not the data itself — just what exists, what format it's in, what domains it covers, how recent it is. When a researcher at one org needs housing data from three cities, they can discover that three other Alliance partners already have it.

Public + Private Dataset Registry

Curate and connect public datasets (census, government open data, academic research) alongside partner-held private datasets. Build the index that makes it possible to answer: "What data exists — public or within the network — that's relevant to child safety in Southeast Asia?" or "Which partners are tracking online hate speech metrics, and how do their methodologies compare?"

Cross-Network Analytics

For organizations that opt into data sharing (with proper governance — this is where frameworks like Better Deal for Data provide the trust layer), build the pipelines that enable cross-org analysis. Aggregate outcome data across similar programs to benchmark effectiveness. Identify patterns that no single organization can see from their own data alone.

AI/ML on Pooled Data

The real leverage: models trained on data from across the network. An NLP classifier for hate speech that's trained on labeled data from CCDH, Fairplay, and CyberPeace Institute — instead of each org building their own from scratch. Donor behavior models trained on anonymized giving patterns from dozens of organizations instead of one. Outcome prediction models that draw on longitudinal data across programs serving similar populations.

This is the data flywheel that no single organization can build alone. The Alliance provides the trust network. We provide the technical infrastructure.

Build Phases

Why This Matters

The Alliance is the connective tissue between 175+ organizations. But right now that connection runs through people, not infrastructure. The internal transformation gives the team the tools to manage a growing network without burning out. The data intelligence service turns the network itself into a platform — where the collective knowledge of 175+ organizations is greater than any single member could achieve alone.

Every other proposal on this site is about building AI/ML infrastructure for one organization. This one is about building it for the network.