The moment I keep coming back to happened outside one of the conference venues in Toronto. It was the second or third day (I don't remember exactly) and our team had stepped aside after a long stretch of sessions. We were tired, a bit overstimulated, and a little stuck.

We'd spent months building a chatbot. We knew why we were there. We knew who we were building for. And yet someone said, almost offhandedly: "Wait, why are we building this chatbot in the first place?"

It wasn't confrontational. No one was calling the work useless. But the question hung there longer than it should have, and I remember realizing we hadn't really paused to ask it in that way before. We had been so focused on how to move forward that we hadn't stepped back to think about whether the thing we were moving toward still made sense.

That huddle outside the conference venue didn't produce a new solution. It produced something more uncomfortable: doubt. And that doubt turned out to be the most important part of the entire REACH experience for me.

Going into the project, I thought I understood what we were doing. We were designing a scalable, tech-enabled intervention to support migrant workers "achieve" financial literacy. In theory, it was clean. Identify a gap, build a tool, measure impact. It fit neatly into the way I had been trained to think about problems: define, optimize, scale.

Looking back, I can see how narrow that frame was.

Some of this was probably my own background showing. As someone trained in computing and systems thinking, my instinct is to reach for structure quickly. Frameworks feel grounding. They give you something to hold onto when the problem is messy. But over time, I started to notice how often we were trying to make reality fit the framework, instead of questioning whether the framework belonged there in the first place.

I still recall a call with Professor Dilip Soman asking if he had recommendations of a framework we could reference in our chatbot design. "Don't look for a framework, go build it." I remember the silence after he said it. I think we were waiting for him to follow up with something, a suggestion maybe, but he just left it there. It was uncomfortable, but also kind of freeing — like we'd been waiting for someone to tell us it was okay to just figure it out ourselves.

Sometimes there really isn't an existing framework—or at least not one that travels well.

That became clearer as we spent more time listening to the workers we were supposedly designing for. Not "listening" in the abstract sense, but actually paying attention to how diverse their situations were, how much knowledge they already had, and how many systems they were already navigating daily. What we saw didn't look like a population waiting for information to be delivered. It looked like people making complex decisions within constraints we hadn't fully accounted for.

Co-creation, in that sense, wasn't about adding features. It was about realizing how much intelligence already existed in the room and how little of it our solution was capturing.

Toronto made this realization sharper. Meeting other Reach teams, what struck me wasn't how polished their work was, but how open everyone was about uncertainty. Rethinking your approach wasn't treated as failure. It was treated as the work.

That feeling culminated during the conference at SMU marking Reach's ten-year anniversary, especially during Joseph Wong's speech. There was one line in particular that I haven't been able to shake since:

"The hardest to reach are often not hard to reach. They are underserved within existing systems."

It sounds obvious when you hear it. But for years, the dominant assumption had been that reaching marginalized populations was primarily a logistical problem—distance, cost, delivery. But Joseph argued that the real barriers were systemic: institutions, mental models, economic assumptions, norms. The issue wasn't access alone. It was trust.

He shared an example from Northern Ontario, where COVID-19 vaccines were successfully delivered to remote Indigenous communities not by optimizing supply chains, but by ensuring the vaccines were administered by Indigenous physicians. Trust, built over time, mattered more than efficiency. The solution worked because it respected history rather than trying to bypass it.

Sitting there, I realized how much of our thinking—not just in this project, but more broadly—still revolves around optimizing the wrong variables.

"We ask how to move from 50% to 70% uptake. We rarely ask how to reach the 30% who are structurally excluded."

That stuck with me. It made me look back at our own work more honestly. Maybe the chatbot wasn't inherently wrong. But maybe it was insufficient. Maybe what migrant workers needed wasn't another interface, but stronger legal protections. Maybe the most meaningful interventions lived upstream: in policy, regulation, employer accountability, access to institutional support.

Those weren't things we could build in a semester. But that didn't make them irrelevant. If anything, it made them more important.

It also made me uncomfortable in a productive way. Because it meant admitting that some of the most pressing problems can't be solved by the tools I'm most fluent in—and that sitting with that discomfort might be more valuable than reaching for the next thing to build.

What Reach gave me wasn't a neat conclusion. It gave me a shift in orientation.

I came in thinking about solutions. I'm leaving thinking more about infrastructure — who builds it, who it's actually for, whether it holds up when trust is thin. I'm also a lot more suspicious now of what we choose to measure in the first place, and what we quietly decide doesn't count.

Most of all, I'm leaving with a deeper respect for listening as a form of work. Not as a preliminary step before building something "real," but as the core of the process itself.

Becoming a REACH researcher doesn't feel like something you finish. It feels like something you carry with you—into future projects, into whatever work comes next. A reminder that the goal isn't always to fix what's broken quickly, but to stay engaged long enough to understand what's actually holding the system together and who it's leaving out.

---

Acknowledgements

This work wouldn't exist without a lot of people who showed up before there was anything to show for it. Thank you to Professor Andrew Koh for the nomination and for backing us even when we weren't sure where the project was headed; to Marin and Joseph for pushing us to think harder about what research like this is actually for; and to everyone in the Reach Alliance community across 2024 and 2025 who made the work feel less isolating than it could have been. Most of all, thank you to the migrant brothers who let us into their lives, their conversations, and their kitchens.