An AI-powered intelligence platform that helps under-resourced nonprofits turn scattered information into clear next steps.
Parsons Capstone Project
7 Months
UXUI Designer
Figma、Cursor、Claude
BACKGROUND
Nonprofits put a lot of effort into expanding their impact. They run campaigns, launch programs, and reach new communities, hoping to drive real change and attract the funding to sustain their work. But the results don’t always match the effort. Programs reach fewer people than intended. Communities don't feel the change. Donors don’t stay engaged.
We assumed the reason was a lack of effective social listening tools. That left nonprofits unable to track conversations and public sentiment across platforms closely enough to know what their communities truly needed. So we dug deep into the problem.
INTERVIEW
We conducted 10 semi-structured interviews, deliberately spread across three levels of the organization: individual contributors doing the day-to-day work, leadership setting strategy, and outside advisors who study the sector from a distance. This helped us understand the problem from multiple angles and test our initial assumptions against real-world experience.
We synthesized these conversations through affinity mapping and thematic coding, looking for patterns across roles rather than isolated complaints. What emerged didn't point to a listening problem at all.
Too many signals, not enough hands
Nonprofits are often expected to track policy shifts, social media engagement, community conversations, and program progress all at once, with too few people to keep up with them all.
Yes, we’d use data, if the data was easily available and wouldn’t take away from others’ responsibilities...No, we are not using data at this time.
Executive Director, HEAL Palestine
Data gets analyzed, but decisions don't always follow it
Individual contributors do the analysis, but leadership makes the call. However, leaders often rely on intuition instead of the evidence that analysis produces.
I believe we aren't able to solve these problems because in the NGO space, people make decisions based on what happens in meetings and previous years' experiences.
Technical Delivery Lead, International Rescue Committe
Impact that isn't felt breaks trust
When communities don't see how their feedback shaped a decision, engagement starts to feel one-directional. Over time, that erodes trust, and people stop offering feedback at all.
They don't feel heard. They feel like things are happening between close doors...rather than leveraging their ideas.
Technical Delivery Lead, International Rescue Committe
Knowledge that leaves when people do
Institutional knowledge in nonprofits is often tacit and person-dependent. When short-staffed teams see high turnover, the context behind past decisions leaves with the person who held it.
COMPETITIVE RESEARCH
No single tool addresses the full picture. Organizations are left stitching several tools together, and still missing pieces.
Most of these tools were designed for corporate use cases. So, nonprofits have to bend their own workflows to fit tools that weren't built for them.
The more capable, integrated tools tend to be the most expensive, beyond what most under-resourced nonprofits can justify spending.
Even where tools do collect and organize data well, almost none analyze it or recommend a next step. They leave the interpretation entirely up to the organization.
IDEATION
VALIDATION
We went back to three participants from the first interview round and asked them to rank six proposed features using a forced-ranking exercise. Since building all six into an MVP wasn't realistic, we needed to understand which three mattered most and whether they would actually address the problems organizations faced. Donor 360, Community Sentiment Intelligence, and Policy & Stakeholder Intelligence consistently ranked highest.
Community Sentiment Intelligence sat in an already crowded space, so building another version of it on its own didn't add much value. We merged it with Peer Benchmarking instead, turning it into something closer to a community: instead of just tracking social sentiment, nonprofits could compare notes and learn directly from organizations facing similar challenges.
Design
With the three features validated, we built Ekko into a working product using Cursor and Claude.
You can also explore it yourself at https://ekko-psi.vercel.app/