FCCF

Applying AI to research that serves people

I served as a volunteer Encore Fellow with Fairfield County’s Community Foundation from September through December 2025 and as a consultant from December 2025 through April 2026. I used AI extensively to support nonprofit research, data analysis and planning. The work addressed two related questions: how to understand the county’s civic organizations more clearly, and what a practical initiative to strengthen social connection might require.

Those questions matter because an organization’s visibility in a dataset, its capacity to participate and its value to a community are different things. I wanted the research to help people make informed decisions about the organizations and communities involved.

My role

I helped define the research scope, co-produce civic-landscape data, document methodology and develop feasibility and implementation materials. I worked with the foundation’s research team on the problems that arose as the work progressed. AI assisted with research, spreadsheet work, synthesis and drafting; my contribution included setting requirements, reviewing results and challenging conclusions that the available evidence did not support.

Making fragmented information reviewable

The civic-landscape work brought together information from nonprofit directories, tax records and other public datasets. The working materials retained source information alongside organization records and included fields for identifying possible duplicates and reconciling website information. We also organized material by issue focus so that the research could be examined from more than one perspective.

That structure made it possible to ask where a record came from, how it related to another record and what still needed attention. A consolidated list becomes more useful when someone can inspect the evidence behind it.

Questioning the classifications

One of the most consequential parts of the work was reviewing classifications and identifier matches. I questioned a pattern in which many organizations had been assigned a partial civic focus, and asked for the evidence behind those conclusions. The review exposed unsupported defaults and labels that overstated the research performed. I also challenged row alignment and requested further verification where identifiers or records appeared questionable.

The methodology addressed missing information as well. A missing classification could call for further research rather than automatic exclusion. Preserving that distinction matters when the purpose is to understand a community: incomplete source data should remain visible as a limitation in the analysis.

Understanding what participation would require

For the social-connection initiative, the stakeholder analysis considered what different organizations might contribute, why participation could be worthwhile for them and what might constrain it. Capacity, access, confidentiality and competing responsibilities all affect whether a promising idea can work in practice.

I used that analysis to develop a more grounded proposal. A partnership has to make sense for the people being asked to sustain it, including organizations whose time and resources are already stretched.

Turning a concept into operating decisions

The feasibility work developed the Chamber of Connection concept into a proposal covering governance, responsibilities, resources, affiliation considerations and phased implementation. It also considered how local readiness should affect the order of activities.

Evaluation was part of that proposal. Someone expressing an intention to build a connection is an encouraging signal, but follow-up is needed to understand whether that intention became action. The measurement approach distinguished those stages rather than treating an immediate positive response as demonstrated impact.

What the work produced

The resulting materials included civic-landscape workbooks, methodology notes, issue-focused research, stakeholder assessments, and a feasibility and implementation proposal. They were developed to support decisions about where the foundation might lead, convene or support existing work.

For me, this project illustrates what useful AI work requires: clear direction, scrutiny of the output and attention to the people who will rely on it. The value comes from helping someone understand a situation well enough to act with greater care.