From AI-Assisted Checking to Collaborative, Evidence-Based Decisioning
AI PreCheck represents a fundamental shift in how development applications are assessed - moving from manual, document-heavy workflows to real-time, AI-driven evaluation.
AI PreCheck 1.0 introduced this shift by enabling instant compliance assessments on standard plan submissions, significantly reducing the time required to identify issues and improving early-stage feedback. However, the process still relied on a centralised human-in-the-loop model, where Archistar performed quality assurance before results were returned - typically within one business day.
AI PreCheck 2.0 removes this bottleneck and evolves the system into a fully collaborative workflow between applicants and government.
Applicants can now upload their submission and receive a detailed AI assessment in under three minutes. They can review, validate, and amend the results, tag supporting evidence, and provide structured commentary - transforming the submission into an evidence-based, AI-assisted package.
This package is then provided to the city, where planners can review both the AI output and applicant inputs, add their own feedback, and make decisions with full transparency.
This evolution introduces a new model for planning assessment:
- Distributed human-in-the-loop validation (Applicant + City)
- Evidence-based completeness instead of document-based checking
- End-to-end collaboration across the submission and assessment lifecycle
The result is faster assessments, higher-quality submissions, and a transparent, auditable decision-making process that aligns with real-world planning workflows.
Comparison
Section titled “Comparison”AI PreCheck 2.0 evolves from an AI-assisted checking tool into a collaborative, evidence-driven decisioning platform - shifting from internal QA to a shared validation model between applicants and government.
| Capability | AI PreCheck 1.0 | AI PreCheck 2.0 |
|---|---|---|
| Turnaround Time | Up to 1 day (AI + Archistar QA) | Under 3 minutes (real-time AI) |
| Human-in-the-Loop | Archistar (centralised QA) | Applicant + City (distributed validation) |
| User Role | Passive recipient of results | Active participant (review, amend, comment, annotate) |
| Submission Model | AI-generated report | AI + Applicant-reviewed + evidence-backed submission |
| Collaboration | Limited | End-to-end collaboration between applicant and city |
| Validation Model | Centralised internal QA | Multi-party validation (Applicant + Authority) |
| Evidence Capture | Limited / implicit | Structured, rule-linked evidence capture |
| Annotations | Basic AI outputs | Smart annotations + user markups |
| Measurement & Checks | Fixed outputs | Scalable, dynamic measurement and validation |
| Transparency & Auditability | AI output only | Full audit trail (AI + applicant + city inputs) |
| Workflow Position | Pre-submission insight tool | Integrated submission and assessment workflow |
| Outcome | Faster initial feedback | Faster, higher-quality, and more consistent decisions |