Innovation Advisory · Framework 14
5 questionsThe Rapid Civic Tech Canvas
Five questions to sit with before building a public tool at the speed of the news cycle that created the need for it. Drawn from the Democracy by Design build.
What piece of the public record will be gone in a week if you do not act?
Rapid civic tech justifies its speed only when the artefact it protects will soon disappear. Protest placards are thrown away by Monday, a leaked WhatsApp screenshot is pulled by Tuesday, and a page is taken down by Wednesday. If the record is not disappearing, use a different framework and build slowly.
Who could this tool hurt, and what protections have you built in before you shipped?
The instinct to ship fast is right, and the instinct to add safety later is wrong. Identify the vulnerable subjects (protesters, whistleblowers, minors, sources), specify the harms (identification, state retaliation, targeting), and put the guardrails into the first commit. Decide on automatic face obscuring, verification of takedown requests and keeping no logs of who submitted what before you buy the domain name.
What will this tool refuse to become, even when asked politely?
A one-line archive of protest posters drifts into a general news outlet, then a commentary platform, then a general political site, unless the refusal is written down. Decide the scope in the first hour and write the refusals into the editorial guidelines. When somebody asks for the tool to widen, point them to the refusals.
How do you care for the record without claiming certainty about it?
Every submission gets a human check and an automated pass (EXIF, metadata, cross-reference against public sources). Publish the chain openly and do not present the archive as a court. An archive that says what it can and cannot verify behaves like a witness. An archive that claims certainty can be used against its own subjects.
Can this ship at the speed of the news cycle that created the need for it, and if not, is it the right project for that cycle?
Some civic tools have to exist within days of the moment they respond to, and others should not be rushed. The two-hour test does not require everything to ship in two hours. It asks whether the moment that created the need will still be present when the tool arrives. If it will have passed, the tool is a memorial. If it will still be there, ship.
The position this canvas takes
Rapid civic tech is a legitimate mode of practice, as serious as the work the slower canvases support. The five questions are the discipline that speed requires: the guardrails, the refusal, the verification chain, and a plain account of what the moment can and cannot support. A tool that answers all five in its first hour is a tool that survives its second week.
If any of the five answers is unclear when you sit down to build, wait. Stay with the question that failed until it is resolved. The speed applies to shipping, and the thinking still takes as long as it needs.