Public consultation is at the heart of urban planning practice in Quebec. It underpins the legitimacy of land-use decisions, structures the dialogue between elected officials, professionals and citizens, and feeds planning documents that commit communities for decades. Yet in many municipalities, the process still relies on methods that struggle to capture representative participation and that generate a volume of data difficult to process rigorously within the usual timelines.
Artificial intelligence opens up concrete ways to ease some of these operational constraints. Not to automate local democracy, but to free urban planners from the most laborious tasks so they can focus on what really matters: listening, analyzing and making informed decisions.
To place these applications in a broader perspective, the page AI and the urban planner in Quebec: how to use it in your daily work offers a useful overview.
Current challenges of public consultation in Quebec municipalities
Before exploring what AI can bring, it helps to name the concrete difficulties urban planning teams face when running a consultation exercise. These are the friction points technology can help reduce.
A participation rate that does not reflect the diversity of the territory
Public consultation sessions generally draw a fairly homogeneous profile of participants: established homeowners, long-time residents of the neighbourhood, people available on weekday evenings. Tenants, workers with atypical schedules, newcomers and young households take part very little, not out of disinterest but for lack of real access to the process.
This reality is well documented in Quebec consultation practices overseen by the Ministry of Municipal Affairs and Housing. Digital tools, some of which now include AI features, make it possible to diversify participation channels by offering asynchronous options accessible remotely, which potentially broadens the range of voices that reach urban planners.
Manual processing of briefs and citizen comments
A public consultation on an urban plan revision or a major development project can generate dozens of briefs, hundreds of online forms and several hours of audio recordings. Summarizing all of these contributions faithfully, in a balanced way and in a form usable for the next steps of the process is considerable work, usually falling to one or two people on tight deadlines.
AI can step in directly at this stage to speed up processing without sacrificing the quality of the analysis.
What AI concretely brings to the consultation process
The applications presented here are operational in 2026 with accessible tools. They require neither a custom-built IT system nor a technology budget beyond the reach of an average-sized municipality.
Automatic summary of briefs and comments
Submitting all the briefs received during a consultation to an AI tool to extract a structured thematic summary is the most directly useful application in this context. The tool identifies the most frequently raised subjects, groups similar positions, isolates minority concerns that still deserve to be documented, and produces a summary report the urban planner can use as a working base.
What would have required two days of intensive reading can be brought down to a few hours, including validating the report produced. The urban planner remains the final reader and the judge of the summary's relevance: AI prepares the material, the professional interprets it.
Detecting recurring themes in citizen contributions
Beyond a linear summary, AI can map the main concerns expressed by participants by grouping contributions according to the themes they address: mobility, densification, green spaces, built heritage, local services. This thematic mapping makes the consultation report easier to write and shows citizens, transparently, that their contributions were read and taken into account as a whole.
For public consultations carried out as part of RCM obligations regarding the revision of land use planning schemes, this type of structured report also meets the transparency expectations of the Ministry of Municipal Affairs and Housing.
Preparing facilitation questions for sessions
AI can also support the preparation of consultation sessions by generating facilitation questions calibrated to the project under review, the profile of the territory and the anticipated issues. A well-built list of questions helps the urban planner facilitating the session cover the essential aspects of the project, restart discussion when it stalls, and maintain a balance between the different themes to be covered.
This is not a task AI does better than an experienced urban planner in every case. But for a professional preparing a session on short notice or on a type of project they know less well, it is a useful starting point that improves quickly with a few adjustments.
Ethical limits that should not be sidestepped
Enthusiasm for these applications should not come at the expense of serious thinking about the associated risks. In the field of citizen participation, these risks have a democratic dimension that makes them particularly important to name.
Bias in the analysis of citizen opinions
An AI tool that analyzes citizen contributions does so using language models trained on corpora that do not necessarily represent the diversity of voices in a Quebec municipal territory. Certain registers of language, certain ways of expressing a position or certain local cultural references may be misinterpreted or underweighted in the summary produced.
This means an AI-produced summary can, unintentionally, over-represent some points of view at the expense of others. The urban planner using these tools must keep a critical eye on the results and read a sample of the original contributions to confirm that the summary faithfully reflects the range of positions expressed.
Transparency with citizens about the use of AI
The question of whether citizens should be told that their contributions will be partly analyzed by an artificial intelligence tool no longer really arises: in a context of fragile institutional trust, transparency on this point is both an ethical obligation and a practical precaution. A municipality that processes citizen briefs with AI without disclosing it risks having the legitimacy of the process challenged if the information comes out afterwards.
A simple mention in the consultation documents is enough in most cases. It reflects an honest approach and can even strengthen the credibility of the exercise by showing that the municipality is using the means available to process contributions seriously.
For a deeper look at these issues, the article on the limits of AI in urban planning addresses bias and professional responsibility in a broader framework.
Where to start for your next consultation
The first step is not technological: it is choosing a precise, bounded task on which to test the tool during the next consultation exercise. Summarizing the online comment forms received after a public session is a low-risk and immediately useful entry point. The task is repetitive, the volume can be high, and the summary produced is easy to check by comparing a few original contributions with the result.
Once this first test proves conclusive, applications can be expanded gradually: thematic analysis of the briefs submitted, help preparing subsequent sessions, drafting the final consultation report. Each additional step builds on the confidence gained in the previous ones.
Are you preparing a public consultation soon and want your team to master these tools before diving in? The artificial intelligence training for Quebec urban planners and municipal teams can be adapted to your schedule and your specific context.
Frequently asked questions
Can AI replace the facilitator during a public consultation session?
No, and the idea should not be seriously considered in a local democratic setting. Public consultation is a space for human exchange where reading the room, managing tensions and being able to reformulate positions are irreplaceable relational skills. AI can prepare the facilitator, equip them and help process the results afterwards. It cannot hold the room in their place.
How can you make sure the summary produced by AI is faithful to the contributions received?
By validating a sample. Once you have an automated summary, reread about ten randomly chosen contributions among those processed and check that the positions expressed appear in the summary. Pay attention to short contributions or those written in an informal register, since they are sometimes captured less well than texts written in formal French. This validation takes little time and provides a reasonable level of confidence in the overall quality.
Is municipal council approval required to use AI in a public consultation?
To our knowledge, there is no specific legal obligation to that effect under the current framework in Quebec. That said, informing elected officials of the use of these tools in consultation processes is sound governance practice. In some politically sensitive contexts, obtaining formal council approval before deploying these tools can also prevent later questions about the regularity of the process.
Do online consultation platforms already include AI features?
Some do, yes. Digital consultation tools used in Quebec municipalities offer automatic comment grouping, trend visualization and summary report generation. The level of sophistication and reliability varies from one platform to another. Before adopting one, it is worth examining how the AI features are documented, what data is sent to the vendor's servers, and to what extent the results produced have been validated in contexts comparable to yours.
Can AI help reach groups of citizens usually absent from consultations?
Indirectly, yes. Digital tools accessible on mobile, available in several languages and usable outside evening session hours can make participation easier for citizens who do not travel to meetings. AI contributes here by making these platforms more intuitive and enabling smoother interactions across different registers of language. But the tool alone is not enough: reaching usually under-represented audiences also requires active outreach that cannot be delegated to an application.
« » »
