☀️ Field-First & Integration Opportunities Brief

1. RFI → Change Order Auto-Link in Buildr+

The Problem: GCs lose track of which RFIs actually spawn change orders. Critical data lives in two disconnected places — the RFI history and the Change Order log — leading to scope confusion and cost disputes.

The Solution: Auto-detect cost and schedule implications in RFI descriptions, then auto-draft linked Change Orders with estimated impact pre-filled. Saves ~2 hours per project connecting the dots.

Implementation: Create a Supabase Edge Function that watches for new RFI submissions. Parse the RFI text and attachments for cost keywords ("budget," "estimate," "quote"). When detected, auto-draft a Change Order with the RFI ID pre-linked, estimated cost, and timeline metadata pulled from the RFI. Present it to the user for final review and sign-off.

• Edge Function: parse RFI text for cost keywords + metadata extraction
• Pre-populate Change Order form with linked RFI, cost estimate, impact summary
• Keep RFI-to-CO relationship in Supabase for audit trails

2. Offline Signature Capture with Fallback in Inspectr+

The Problem: Field inspectors lose connectivity mid-signature on transmittals and lien waivers — a critical workflow interruption that stalls project closeout.

The Solution: Add a local buffer for signature strokes, queue them for sync when online. Fallback: capture photo of signature + timestamp if touch input unavailable (e.g., on older devices or with gloves).

Implementation: Use Flutter's signature package to buffer strokes locally. Store the stroke vector data in SQLite with a sync queue (status: pending/synced). On reconnect, validate signatures against Supabase realtime. Fallback: add a "photo signature" option that captures timestamp + photo + inspector metadata.

• Gesture detector captures touch strokes, queues locally
• SQLite queue: (signature_id, vector_data, document_id, status)
• Supabase validation: confirm signature is in queue on reconnect
• Fallback: photo + timestamp for gloved work or older devices

3. Smart Document OCR & Auto-Classification in Buildr+

The Problem: Teams upload mixed batches of construction documents (meeting minutes, invoices, GC notes, RFI responses) — then manually sort and tag them. Massive busywork that delays document management.

The Solution: Supabase Edge Function + Claude Vision API auto-sorts docs into Submittals, Change Orders, Meeting Minutes, etc. Reduces manual tagging by 80%.

Implementation: Create a Supabase Edge Function that triggers on document upload. Call Claude's Vision API (or GPT-4V) to analyze the document image. Extract: document type, date, parties involved, and key metadata fields. Auto-populate the doc_type and metadata in Buildr+. User can override if needed.

• Edge Function: on_document_upload → call Claude Vision API
• Extract: doc_type (RFI, Change Order, Submittal, Meeting Minutes, Invoice, etc.)
• Fill metadata: date, parties, cost (if applicable), project phase
• Reduce manual tagging from hours to seconds per batch

Why These Three?

All three address construction industry pain points and leverage modern integration patterns:

• RFI-to-CO Auto-Link → Faster cost tracking & audit trails (Buildr+ core workflow)
• Offline Signatures → Reliable field workflows even in poor connectivity (Inspectr+ critical path)
• Smart OCR → Automate the busywork, let PMs focus on strategy (Buildr+ document hub)

Each integrates directly with the Flutter + Supabase stack. No new infrastructure — just smart use of Edge Functions + modern APIs.