https://www.linkedin.com/pulse/what-could-ai-do-community-engagement-martijn-de-waal-zkhgc?utm_source=share&utm_medium=member_ios&utm_campaign=share_via

What could AI do for community engagement?

From Nina Mackens presentation: speed doesn't replace trust. If people don't trust a participation process, they will find other ways to voice dissent

What could AI do for community engagement? This was one of the central themes emerging from the Community Engagement Summit (https://cesummitseries.com/) I attended in Sydney this week. It was a great meet-up of around 100 community engagement professionals, mostly from across the Greater Sydney area. Meanwhile, discussions arose about the essence of the profession: optimizing participatory processes, or creating meaningful inefficiencies to invoke deeper engagement?

It struck me that many participants found themselves caught between two forces. On the one hand, they expressed a growing pressure to take part in the “compliance theatre” demanded by city governments eager to push urban development plans forward efficiently, with as little friction as possible. On the other, they expressed their desire to follow their professional ethics and intrinsic motivation to engage meaningfully with local communities: building trust between governments and residents, giving people a real voice, and ensuring their interests are genuinely represented.

Article content

From Nina Macken's presentation: engagement needs legitimacy and trust, and it takes time to build that up

Could AI come to the rescue and restore their professional values?

Speakers from various software companies argued that it could. True, as one of them acknowledged, AI cannot knock on doors (at least not yet—though it could probably already conduct qualitative phone interviews), nor can it chat with passers-by at pop-up participation booths. But, the argument went, these moments of human-to-human engagement make up only 30–40% of the work. Much of the rest is procedural: planning, desk research, analysing results, drafting reports. In particular, analysing qualitative data currently consumes a great deal of precious time. In all these phases, AI could help—freeing up time that could then be spent on actual engagement.

Several AI-assisted tools were introduced:

• Oliver Young presented Scopomap, which seems particularly useful in the research and exploration phase. Its chatbot, “Bruce,” is trained on urban data and can quickly provide insights into demographics and neighbourhood characteristics, producing stakeholder analyses and community personas almost instantly.
• Tom Workman introduced Converlens, a tool focused on qualitative data analysis. What stood out was its ability to combine hard-coded data (for example, demographic variables) with qualitative quotes and insights. This is something standard LLMs are currently notoriously bad at. Ask them for quotes from a specific demographic, and chances are they will fabricate both the person and the quote. The vibe may be right—but that’s not good enough for an official report.
• Don Sharples demonstrated Social Pinpoint, which offers multiple tools to collect and analyse data, combining inputs from different channels into integrated insights.
• Dan Ferguson represented Communiti Labs, offering similar capabilities, including the option to automatically report outcomes back to participants.
• Charles Connell pitched CE Canvas, again with overlapping features, including report writing.

The presentations were too short to fully appreciate the nuances of each system, but their promises were broadly similar. These tools are trained on data specific to cities, development and community engagement—rather than the internet at large—and therefore claim to deliver more reliable and context-aware results. They also include audit trails, allowing users to trace insights and quotes back to original sources, reducing the risk of hallucinations.

Were community engagement professionals enthusiastic about these prospects? Yes—and no. Especially in organisations with limited resources, it is clearly attractive to make data collection, analysis and reporting more efficient.

But this raises a deeper question: what exactly do we mean by efficiency?

From a systems perspective, efficiency means doing more with fewer resources. From a professional perspective, that is not necessarily the goal. For community engagement practitioners, the report itself is not the endpoint. The real goal is building trust and involving communities in democratic processes. That may require what Eric Gordon and Gabriel Mugar have called “meaningful inefficiencies”: the time-consuming work inherent to democracy. Inefficient from a management point of view, perhaps—but essential for creating meaningful interactions between citizens, and between citizens and governments.