How Co-Designing with AI Can Speed Up Community-First Design

Last year, I found myself at a familiar crossroads for many service designers — my team had access to an off-the-shelf product that almost met our needs, but not quite.

I was working with the Kativik Regional Government to  modernize water and waste delivery in Northern Quebec. The service delivery drivers and operators work in demanding, time-sensitive conditions shaped by limited connectivity, multilingual needs and a deeply relational context. 

Through research and co-design with drivers and operators, a gap became clear. Their workflows, language needs and physical environment are highly specific and deeply relational. A generic tool couldn’t adapt to how the work actually happens on the ground.

Knowing that is one thing. Making the case for a different approach is another.

Building a custom application requires time, talent and trust. Before our team could commit, I needed to show not only where the chosen off-the-shelf product created friction, but how a tailored solution could better support the people who would use it every day. The challenge wasn’t convincing through words — it was creating something concrete enough to react to.

Why AI works well for early-stage prototyping

At our 2025 all-staff retreat, Code for Canada ran an internal AI hackathon.

We worked in small teams, with tight timelines, on a mock project we had to take end-to-end using AI. The goal wasn’t to build something perfect. It was to explore what was possible.

That week changed how I thought about AI. I stopped seeing it primarily as a productivity tool and started seeing it as a way to explore ideas early — to prototype possibilities before committing significant resources.

How to prototype with intent

I used a combination of tools to bring a concept to life:

• Lovable and Replit to generate a functional, clickable interface
• ChatGPT to help structure prompts, test interaction patterns, and synthesize existing research
• My product knowledge and community research to ground everything in lived realities

To keep the work anchored in what drivers and operators shared with us, I structured prompts using a simple framework: Context, Task, Format, Constraints.

Those constraints mattered. They reflected what we heard repeatedly in research:

• Limited and unreliable connectivity
• Multilingual working environments
• Unpredictable daily conditions
• Different levels of comfort with technology
• Strong community values around care, respect, and clarity

How prototypes help secure buy-in

When I shared the prototype with the development team, the conversation changed.

Instead of debating abstract trade-offs, we could see — together — where the commercial product fell short, how rigid templates clashed with real workflows, and what became possible when local context led the design.

The question shifted from “Is a custom solution worth it?” to “How do we build this responsibly?”

That shift mattered. It created shared understanding faster than any slide deck could. 

AI as a companion to human-centred work

AI helped accelerate the translation of community insight into a tangible experience. But it didn’t replace the work that mattered most.

Discernment stayed central throughout the process. AI can generate polished output quickly — sometimes too quickly. Research rigour, cultural awareness, and critical thinking were essential to deciding what to keep, what to challenge, and what to discard.

The value wasn’t in what AI produced on its own. It was in how it supported a process grounded in listening, testing, and respect for the people most affected by the outcome.

Using AI to speed alignment in community-first design

This approach didn’t just help me prototype faster. It helped me communicate more clearly, align our team earlier, and advocate for a solution that reflects the realities of the people we’re designing with.

When other designers ask whether AI has a place in public-sector or community-centred work, this is what I share:

• Use AI to express possibility, not override context
• Use it to speed alignment, not skip engagement
• Use it to support people, not erase nuance

When used thoughtfully, AI can help bridge the gap between insight and implementation — making it easier for teams to design services that truly reflect the people they’re meant to serve.

Inspired? Let's build something together.

If you’re working in the public or nonprofit sector and navigating how to introduce AI without losing trust or context, Code for Canada partners with teams to explore these questions in practice. Get in touch today.