About
About CAIS
CAIS (Construction AI Solutions) is an AI-powered platform designed to automate the entire construction project lifecycle. Starting with Quantity Take-Off, we are building a comprehensive suite of intelligent agents that handle the most time-consuming and error-prone tasks in construction management.
Our mission is simple: replace days of manual work with minutes of AI-powered processing, while maintaining the accuracy and reliability that construction professionals demand.
Technology
How the AI Works
Dual-Path, Two-Pass Architecture
The Building QTO Agent uses a sophisticated dual-path, two-pass architecture powered by Google Gemini. Two independent extraction paths — visual counting and text extraction — are merged and validated through confidence gating and 15 sanity checks before producing the final BOQ report.
PDF Classification
The system analyzes your PDF to detect whether drawings are vector (digital CAD), scanned, or mixed. This determines the optimal processing strategy.
Image Path (Visual Counting)
PDF pages are rasterized at 300 DPI into tiles. Gemini Flash performs pure observation: See, Identify, Extract. Each structural and architectural element is visually counted.
PDF Path (Text Extraction)
Direct PDF text extraction captures dimension chains, schedules, notes, levels, and architectural data. Excels at tabular data like door/window schedules and finish matrices.
Dual-Path Merge
Results from both paths are merged: visual counting from Image path, dimensions and schedule data from PDF path. Per-element audit trail records which path contributed what.
Confidence Gating
Each extraction result gets a confidence score. If overall confidence falls below 70%, the system escalates to Gemini 2.5 Pro for a high-accuracy recount.
BOQ Calculation (Pass 2)
Full 9-section BOQ calculation using protocols A through U: structural, masonry, floors, walls, ceilings, doors/windows, waterproofing, insulation, and miscellaneous.
Sanity Checks
15 post-validation checks catch impossible values, missing dimensions, unit mismatches, and architectural inconsistencies before final output.
Excel Report
Professional 3-sheet Excel workbook: Sheet 1 (BOQ items with 9 color-coded sections), Sheet 2 (summary), Sheet 3 (dual-path audit trail).
By the Numbers
Platform Statistics
What Sets Us Apart
Key Differentiators
Dual-Path Extraction
Unlike single-method tools, CAIS runs both visual counting (image path) and text extraction (PDF path) simultaneously, then merges results for superior accuracy.
Self-Correcting AI
Confidence gating automatically escalates uncertain results to a more powerful model. Independent recount provides a second opinion on visual counting.
Full Building Scope
9 BOQ sections covering structural, masonry, finishes (floors, walls, ceilings), doors/windows, waterproofing, insulation, and miscellaneous items.
Transparent Audit Trail
Every BOQ item includes its extraction source, dimension origin, and a dual-path comparison showing what each extraction method found.
The Vision
Automate the Construction Lifecycle
CAIS is building toward a future where every phase of construction management — from initial quantity take-off through scheduling, pricing, and project coordination — is powered by specialized AI agents working in concert.
Quantity Take-Off
From drawings to BOQ in minutes, not days. Currently in production with the Building QTO Agent.
Scheduling
Automatically generate construction schedules with logic links, resource leveling, and critical path analysis.
Pricing
Intelligent cost estimation with market-based rate build-up, regional cost libraries, and bid analysis.
Project Management
Conversational AI for project stakeholders: ask questions, make decisions, coordinate teams in natural language.