Overview
We built an operations intelligence platform that consolidates event execution scoring, labor and timesheet tracking, vehicle telematics, and profitability data into structured dashboards powered by AI orchestration. We partnered to wire data incrementally from existing modules, enforce tenant-scoped migration discipline, and deliver continuous QA across every phase.
Challenges
- Operational information was scattered across modules with no unified way to score how well an event was actually executed, financially and operationally.
- Dashboard visibility was fragmented, leaving leaders without a single consolidated view of operations, delivery, leads, events, and customer satisfaction.
- Labor cost was never linked to the work that generated it, so timesheet and task data could not be connected to events, farm, shop, order, inventory, or volunteer workflows.
- There was no reliable way to capture task-completion timing, so punch and task performance went untracked across employees and jobs.
- Task verification was missing, leaving no proof that work had been completed correctly or on time.
- Profitability was effectively invisible because flower, material, labor, delivery costs and revenue were never tied together per event or per product.
- Migrating historical operational data carried real risk of duplication, partial imports, and cross-tenant leakage without validation or reconciliation.
- Vehicle telematics for mileage, engine alerts, idle time, and arrivals could not be trusted until vendor API access, authentication, and data format were confirmed.
- Reviews, venue knowledge, and internal communications were ad hoc, with no AI layer monitoring task completion, issues, or surfacing operational suggestions.
Solution highlights
- Built an event execution KPI scorecard spanning event snapshot, financial performance, operational execution, production accuracy, customer experience, internal debrief, a final health score, and monthly summary.
- Delivered an AI operations dashboard surfacing retail orders, delivery statistics and map, lead sources, event statistics, customer satisfaction summaries, and location filtering.
- Integrated labor and timesheet tracking from legacy operations with labor cost tracking across farm, shop, task, order, inventory, volunteer, and event workflows.
- Added punch in/out task tracking with employee and task time entries, plus individual and aggregate labor performance reporting.
- Implemented event and proposal item to task assignment with employee-class-based assignment rules and image upload for task verification.
- Built AI-driven schedule suggestions and duration learning that improve over time using events, harvests, orders, inventory, and volunteer days.
- Wired vehicle telematics ingestion for mileage, engine alerts, idle time, speed, arrivals and departures, with vehicle repair integration for engine codes and alerts.
- Delivered real-time profitability tracking and per-event, per-product aggregation across flower, material, labor, and delivery costs against revenue.
- Created an AI orchestrator covering schedules, verification, performance scores, customer feedback, suggestions, job descriptions, and replacement suggestions, plus AI-generated venue profiles and review automation.
- Implemented tenant-scoped JSON data migration with validation, idempotency, and reconciliation reports, backed by a QA workstream across unit, integration, tenant-command validation, UAT, and release regression.
Outcomes
- Leaders can now score event execution against a single structured scorecard instead of piecing together scattered operational information.
- Operations, delivery, leads, events, and customer satisfaction are visible in one consolidated dashboard with location filtering.
- Labor cost is tied directly to the tasks, events, and timesheets that drive it, making workforce spend traceable across every workflow.
- Task completion is now timed and verified through punch tracking and completion image uploads, giving trustworthy evidence of work done.
- Profitability is visible in real time and aggregated per event and per product, so margins are clear where the data exists.
- AI schedule suggestions and duration learning sharpen planning over time using real event, harvest, order, inventory, and volunteer signals.
- Historical data migrates safely with validation, idempotency, audit logs, and reconciliation reports, protecting tenant isolation and data integrity.
- Vehicle telematics and engine alerts feed operations only once vendor access is confirmed, keeping the data dependable rather than speculative.
- Continuous QA on every phase and two-developer parallel execution compressed the timeline while keeping quality and release regression coverage intact.