Ontario permit counters are under pressure from two sides. Housing targets keep climbing, and reviewer capacity isn't growing. Clariti builds permitting and community development software for local governments. Its AI pre-check tool, CivCheck, helps applicants fix their files before staff see them.
Applications arrive incomplete, and staff send them back for the same fixes. The old loop runs like this. An application is filed, waits in the queue, gets comments, and goes back to the end of the line. Each trip costs weeks. Most trips trace back to gaps the applicant could have closed in minutes.
Applicants run the pre-check before they file, so staff see only review-ready files. CivCheck sits between the applicant and the permit system. The applicant uploads plans and documents. CivCheck checks for missing items, such as a north arrow or a professional stamp, and flags possible code problems with the code section cited. The applicant corrects the file and runs the check again. Only then does the file enter the review queue.
Most applicants skim checklists, or never open them. CivCheck turns each checklist into guided steps, so the applicant confirms every item before filing.
Staff make the call and CivCheck shows a finding only when it is confident, and the applicant confirms or corrects it. Reviewers check the results before any determination. They can accept, change, or override a finding, and each flag points to the code it relies on.
AI should assist, not decide. It should not replace licensed reviewers, act as a black box, or approve applications on its own. Permit decisions affect public safety, so people keep final authority.
CivCheck asks who is applying, and the answer shapes which checks run. A licensed architect or engineer gets code and bylaw checks. A homeowner mainly learns what to submit. A developer on a larger project can add code checks when the project needs them.
The applicant picks a property and answers a short set of questions about the project. CivCheck checks the property against GIS layers, including flood zones, and suggests a project scope from the uploaded plans. The applicant confirms or changes that scope.
Each check shows where an item sits on the plan and why staff look for it, with a link to the bylaw beside it.
CivCheck also checks documents against the file. A study that cites a different address gets flagged before filing.
Before filing, the applicant sees a readiness score. Some departments require a minimum score. Applicants can override an item with a stated reason, and reviewers still decide the gray areas.
Each code check states why staff look for the item. The sample demo runs four checks: floor area ratio, smoke alarms, building height, and setbacks. Each check links to the bylaw and ends in a likely-compliant result. A human reviewer still makes the final determination.
Staff review what the applicant confirmed, so they do not repeat the work. Intake and plan reviewers check the work, re-measure where needed, and make the call. A floor area ratio check shows the calculation and the target it must meet. Staff can send correction notices from the same screen, and managers can track volume and review measures on a dashboard.
Many departments start with completeness: documents, stamps, north arrows, and scale bars. They add code checks for specific project types later, once the team trusts the output. Building, planning, and IT each own part of the checklist and part of the system.
Departments already measure the first completeness check. Measure the second resubmission too, because every review cycle adds weeks.
Bring your building, planning, and IT leads to the first conversation. Come with your technology stack, your staffing levels, and your number of plan reviewers. Those three answers show how much a pilot can take on and where the first checks should go.
Book a demo to see the applicant and staff views for yourself. Then watch the full session, including the live demonstration here.