AI for community engagement
CE Canvas, Copilot and ChatGPT: why engagement needs its own AI assistant
Copilot, ChatGPT and Claude are good general assistants. Community engagement needs more than a good model: the context of the project, the method of engagement practice, sources you can check, and a person signing off. That is what the CE Canvas AI assistant adds.
The short answer
Use both, for different jobs
Use Copilot or your organization’s approved chatbot for general writing. Use the CE Canvas AI assistant for the engagement itself, because it works from the project record, follows engagement practice, shows the source of every finding and waits for a person to sign off. The model is not the difference; the context, the method and the evidence trail are.
What engagement needs
Five things a general chatbot does not do for engagement work
General assistants are built to help with anything. Engagement work is judged after the fact, under scrutiny, so it needs an assistant built for that.
It works inside the project
The assistant works from the engagement record: the project brief and decision, the stakeholder map, the activities, the results and the documents attached to the project and to your organization. A general chatbot only knows what someone pastes into it, every time.
It knows engagement practice
It is built for engagement work: it holds plans to the level of engagement your project has set on the IAP2 spectrum, calibrates suggested stakeholders to the decision, and flags gaps such as a level that does not match the decision or missing context.
It shows its sources
Every finding lists the documents, notes or results behind it, and each answer is marked grounded, partly grounded or interpretive. Answers that cite a source the assistant was not given are rejected before you see them.
People sign off
Drafts of plans, documents and reports are held as previews until someone on your team applies or discards them. The assistant does not publish, send, delete or decide anything.
It asks rather than guesses
When a project has too little context to draft from, it asks for the decision, the people affected or a document to work from, instead of producing a confident plan from nothing.
In the product
What it looks like in practice
Every point cites its evidence
Asked which issues are likely to be contentious, it answers point by point with the source for each and lists what it drew from.
A draft, held until a person commits it
The assistant drafts a plan section in place and marks it as not saved. Discard or replace are the only actions.
It asks rather than guesses
On a project with no brief or materials, it declines to draft and asks for the decision, who is affected, or a document to work from.
Compared
A general chatbot and the CE Canvas AI assistant, side by side
- What it knows about your projectGeneral chatbot: What you paste into the chat, each timeCE Canvas: The project record: brief, decision, stakeholders, activities, results and attached documents
- Engagement methodGeneral chatbot: General best practice, whatever the prompt asks forCE Canvas: Built for engagement: the IAP2 level set for the project, stakeholder calibration, practice checks
- SourcesGeneral chatbot: Sometimes, and not always checkableCE Canvas: Every finding lists its sources, with a grounding check on each answer
- Analyzing submissionsGeneral chatbot: Summaries of whatever text fits in the chatCE Canvas: Topics clustered from every response, each linked to the comments behind it, promoted into themes by your team
- ReportsGeneral chatbot: A draft you assemble and check by handCE Canvas: What-we-heard drafts scoped to each phase, with confidence and sources shown
- Sign-offGeneral chatbot: Whatever you copy out of the chatCE Canvas: Drafts held as previews until a person applies them
- Consistency across the teamGeneral chatbot: Each person writes their own promptsCE Canvas: Shared playbooks, templates and organization documents for every project
- Record of what happenedGeneral chatbot: A chat history in one person’s accountCE Canvas: Each assistant action, and each draft applied or discarded, is recorded against the project
Enterprise versions of Copilot and ChatGPT offer strong data protections, and your IT team may already approve them. The comparison here is about fit for engagement work, not about which tool is safer.
Governance
Built for public organizations
The same questions your IT, privacy and procurement teams ask of any AI tool, answered for engagement work.
- Customer documents, submissions and engagement data are not used to train models.
- Models come only from approved providers, Anthropic and OpenAI, through their commercial APIs.
- Your organization’s data is hosted in your country. AI requests may be processed outside it, under contractual safeguards with the model providers.
- Single sign-on with SAML, and each organization’s data kept separate from every other.
- The assistant does not publish, send, delete or decide. What it will not do is documented.
Questions
Common questions about Copilot, ChatGPT and CE Canvas
Can we just use Copilot or ChatGPT for community engagement?
For general writing, yes: drafting emails, tidying a paragraph, summarizing a meeting. For the engagement itself, a general chatbot starts with no knowledge of the project, the decision or your engagement practice, and its answers are hard to trace back to the submissions they came from. CE Canvas does that engagement-specific work inside the project record, with sources shown and staff signing off. Most teams use both.
Can we just use Copilot or ChatGPT?Is it safe to paste submissions into ChatGPT?
It depends on the product and the settings your organization has approved. Enterprise versions offer stronger data protections than consumer accounts. The larger problems are accuracy and defensibility: themes drawn from pasted text cannot easily be traced back to the submissions behind them when a decision is challenged.
The risks of generic AI in engagementDoes CE Canvas use the same AI models?
CE Canvas uses large language models from approved providers, Anthropic and OpenAI, through their commercial APIs, where customer data is not used for training. The difference is not the model. It is the project context, the engagement practice, the source checks and the sign-off built around it.
Is our data used to train AI models?
No. Documents, submissions and engagement data in CE Canvas are used only to produce your outputs and are not used to train models.
Security, privacy and responsible AIDoes CE Canvas replace Copilot?
No. Keep Copilot or your organization’s approved assistant for general work. CE Canvas is for planning, delivering, analyzing and reporting on engagement, where the context, the method and the evidence trail matter.
For the wider picture, read the guide to AI in community engagement.
See the assistant work on one of your projects
A guided pilot runs one real project end to end, with your framework, your documents and your team signing off.