Work

Engagements where AI-assisted operations changed the conversation

Anonymised scenarios from recent PropertyForge AI projects. Outcomes depend on implementation discipline — we document what was designed and human-verified, not what was guaranteed.

Case 01 · Multi-family operator · Burnaby

Rent roll in a spreadsheet nobody trusted

A 340-unit operator reconciled vacancy across three files while maintenance tickets arrived through tenant portal, email and phone. The rent roll in the shared spreadsheet ran everything — until weekly reports showed completion rates that contradicted the maintenance log and nobody could explain the gap before Monday committee.

PropertyForge mapped intake-to-close workflows, used AI-assisted portfolio analytics to surface duplicate data-entry points, and designed a single-queue maintenance model with priority tiers aligned to BC Residential Tenancy Act response expectations. Human-in-the-loop verification sat on every emergency escalation suggestion. We delivered a control matrix, fairness checks on triage logic and a dashboard spec tracking mean time to assign and mean time to resolve.

Engagement fee: C$6,400. Timeline: twelve business days from kickoff to final walkthrough. The client implemented the queue structure internally over six weeks; we remained available for two review sessions during cutover. PropertyForge did not guarantee reduced ticket volume — only a documented target state.

Client operations review session with AI-assisted workflow diagrams on screen

Client review · workflow walkthrough

Vancouver residential portfolio buildings under management

Residential portfolio · Lower Mainland

Case 02 · Institutional holder · Vancouver core

Document intake that missed underwriting

An institutional owner with residential, retail and parkade components received board packs that took eight days to compile. Lease abstracts and insurance certificates lived in shared drives with inconsistent naming. Capex approvals bypassed the underwriting support view entirely because document automation did not exist — only heroic weekend reconciliation.

We engineered document intake routing with machine-learning classification suggestions routed through human verification before filing. Portfolio analytics defined system-of-record for each KPI, refresh cadence and exception-validation rules. Fairness review covered any screening-adjacent fields in tenant documentation workflows. Deliverables included entity-relationship documentation, a KPI dictionary and acceptance tests for the first monthly close under the new structure.

Engagement fee: C$14,200 including two stakeholder workshops. The client's IT contractor implemented integrations over three months using our specifications. PropertyForge did not guarantee NOI improvement — the deliverable was a buildable reporting and document architecture.

Case 03 · Regional operator · Fraser Valley

Maintenance tickets the leasing desk answered forty times a day

A regional operator with twelve buildings fielded identical maintenance questions through leasing staff because tenants did not trust the portal status view. Work orders aged in a queue that had no link to the rent roll's unit register — so priority calls bypassed triage entirely and emergency vendors were dispatched without audit trail.

PropertyForge designed property-management automation linking unit records, ticket status and tenant communication templates. AI-assisted prioritization suggested urgency ranks; operations engineers enforced human-in-the-loop gates before auto-notifications sent. We ran fairness checks to ensure prioritization did not systematically delay certain buildings or unit types. The blueprint included RACI assignments, integration notes and a runbook for parallel operation during cutover.

Engagement fee: C$8,200. Completed in eleven business days. Implementation remained the client's responsibility; our scope ended at engineer-reviewed documentation with two post-launch review sessions.

Case 04 · Developer transition · Surrey

Predictive maintenance after the boiler failed

A developer-holding company missed preventive maintenance signals because work-order history, asset registers and occupancy forecasts lived in separate spreadsheets. A boiler failure the week after a clean quarterly report exposed how little the rent roll and maintenance queue actually talked to each other.

We built predictive maintenance and occupancy forecasting models with confidence bands and documented failure modes — estimates requiring human review before priorities reached field staff. Document automation gates mapped warranty folders to Strata Property Act checklists. Senior data scientists signed off on model assumptions and retraining cadence. Deliverables included swimlane diagrams, a migration-ready runbook and fairness notes on any tenant-facing response-time impacts.

Engagement fee: C$9,600. PropertyForge made no uptime or cost-savings guarantee — only a verified blueprint and monitoring spec the client chose to implement.

Discuss your portfolio question