How it works

Every submission assessed by AI PreCheck moves through the same workflow, regardless of how the platform is deployed. The workflow transforms drawings, documents, and project information into structured findings that can be reviewed by applicants, municipal staff, or both, depending on the selected integration model.

The assessment workflow

Stage 1, Submission Intake: Applicants provide the project information and documents required for assessment., Project and contact info, Drawing and document upload, Submission metadata. Stage 2, AI Interpretation & Extraction: AI converts drawings and documents into structured assessment inputs., OCR and plan classification, Dimension extraction, Annotation recognition, Structured data output. Stage 3, Submission Readiness Assessment: Check whether the submission is complete, legible and ready for assessment., Completeness checks, Missing-item detection, Drawing legibility checks, Readiness score. Stage 4, Compliance Assessment (Deterministic): Deterministic, not AI-driven. Stage 5, Explainable Findings & Evidence: Turn assessment results into traceable findings reviewers and applicants can understand., Pass / fail / attention status, Relevant code clauses, Measurement evidence, Drawing overlays, Applicant guidance. Stage 6, Reviewer Collaboration Workspace: Reviewers resolve exceptions, add context and manage the final review workflow., Confirm or override findings, Add comments and markups, Assign escalations and tasks, Maintain audit trail. Stage 7, Decision & Reporting: Publish the outcome and capture operational signals for improvement., Decision outcome, Reports and exports, Review performance metrics, Improvement signals. Continuous improvement loop feeds back into every stage.

1. Submission intake

A submission is created and project information, drawings, and supporting documents are provided to PreCheck. Depending on the deployment model, this may happen through a municipal permit system, a permit software platform, or directly through Archistar.

2. AI interpretation and extraction

PreCheck interprets the submitted drawings and documents, extracting the information needed to assess compliance and submission readiness. This includes identifying relevant project attributes, dimensions, annotations, and supporting information.

3. Rule selection

Based on the project information and municipality configuration, PreCheck determines which zoning, building code, and readiness rules should be applied to the submission.

4. Compliance assessment

PreCheck evaluates the submission against the applicable rules and generates findings. Each finding is linked to the rule, evidence, and supporting context used in the assessment.

5. Report generation

The assessment results are assembled into a structured report that allows findings to be reviewed, validated, and acted upon. Findings are grouped by category and supported by evidence and visual context where applicable.

6. Review and correction

Depending on the deployment model, findings may be reviewed by applicants before formal submission, by municipal staff during permit review, or by both. Findings can be validated, annotated, addressed, or escalated as required.

7. Decision and submission progression

Once findings have been reviewed and any required corrections have been made, the submission proceeds through the municipality’s normal permit review process. PreCheck supports the review process but does not make permitting decisions.

Continuous improvement

AI PreCheck improves over time through reviewer feedback, applicant behaviour, operational analytics, and ongoing rule and model refinement. This helps ensure assessments remain aligned with municipal requirements and real-world permitting practices.

What happens next?

The workflow described above is common across all deployment models. The way users interact with the workflow depends on how PreCheck is integrated into a municipality’s permitting ecosystem.

• Learn about the available deployment models in the Integration Guide
• Explore applicant and reviewer workflows in AI PreCheck 2.0