What We Do
The non-profit sector is being asked to adopt AI quickly and responsibly — often by funders, sometimes by boards, increasingly by the people they serve. The hard part isn't usually the technology. It is figuring out which decisions warrant AI, which workflows are ready, which staff need what kind of support, and how to do all of that without compromising the values that brought the organization into being.
We bring the rigor of product management to that question. We start with the mission and the actual work. We design adoption strategies that support outcomes — the goal is to make a measurable difference that can be defended to key stakeholders, regulators, or the general public.
How We Help
We are deliberate about treating AI transformation as an organizational change problem first and a technology problem second. Technology adoption is not a goal, it is a tool in a toolkit for achieving goals or furthering the mission of an organization. Our principals' research on non-profit technology adoption and human-centered design has received tens of thousands of downloads and been cited more than 1,800 times across computer science, management, and health journals — the same fields many of our clients are operating in.
Representative Work
Working Paper is a small firm by design. The work below reflects what you're hiring when you hire us — engagements led at Working Paper, and the senior careers our principals brought with them.
Strategic work on the Robert Wood Johnson Foundation's Ask RWJF product. We created and executed strategies for deploying generative AI to advance health policy outcomes, identifying the opportunity for high-impact product work using AI to help people get specific answers based on the Foundation's own quality data. We brought in Miso.ai to implement, and now you can visit RWJF Answers to see how a non-profit's data, with the right strategy and product thinking, can make an impact.
National research program for American schools. We lead AI adoption and digital transformation work for a foundation that is coordinating a national research program to transform American schools — helping translate ambitious philanthropic goals into shippable infrastructure and evaluation frameworks. We delivered these via an MCP server, meaning teachers could get the precise information they needed, and nothing more, simply by asking ChatGPT, Claude, Gemini, or their own favorite AI tool.
AI reliability benchmarking at MLCommons. We lead product, program management, and policy strategy for AILuminate, MLCommons' AI reliability benchmark. Our remit includes the v1.0 Jailbreak Benchmark, multimodal extension, the continuous prompt stewardship system, and the consortium governance artifacts (Responsible Disclosure Policy, Taxonomy Maintenance Process) that make benchmark releases legible to industry, regulators, and the public. The work spans Chatham-House-rule convenings of frontier-model labs, taxonomy stewardship, and the operational machinery a credible benchmark requires.
Strategic investment advising. We advise a foundation on which high-value strategic investments to make. We also lead some of the projects born of those investments, in support of the broader philanthropic and grant-making ecosystem.
Founding a non-profit from scratch. We launched Digitalis Commons, a new non-profit organization, as part of leading global communications and product work for Digitalis Ventures. In addition to supporting contributions to the digital commons, the non-profit was structured with the strategic goal of measurably better deal flow and broader visibility in the healthcare community.
Public Interest Tech Education. Our principals have taught graduate level courses at the Georgetown McCourt School of Public Policy, The University of Cambridge, The University of Edinburgh, and The University of Washington. Courses have covered topics such as using product management skills to improve policy development, delivery, and iteration as well as human centered design in real-world organizational settings. This academic rigor and discipline shapes how we work with non-profit leadership, program officers, and grantees.
Most importantly, we don't ship decks. We ship actual solutions that are co-developed with our clients, meaning they're not left with an implementation gap. The result is AI adoption that is mission-aligned, evaluable, and durable enough to survive the next funding cycle.