AI-Powered Bill of Quantities Analysis: From 200 Positions to Structured Overview

What Is AI-Powered BoQ Analysis?

A bill of quantities (BoQ, or Leistungsverzeichnis/LV in German) is the heart of every construction tender. It describes position by position what services are expected and in what scope. A typical BoQ for a medium-sized construction project contains 200 to 500 positions — of which only a fraction is relevant for any specialized supplier. AI-powered BoQ analysis automates the relevance assessment process: the system reads the entire BoQ, identifies the positions relevant to the specific supplier and categorizes them by product type, requirement profile and complexity. The result is a structured overview that immediately shows sales and costing: which positions are relevant, what exactly is required, and where is there a need for clarification. This process works with GAEB files (X81, X82, X84) as well as PDF bills of quantities and scanned documents.

The Problem: Hundreds of Positions, Manual Matching Does Not Scale

Manual analysis of a bill of quantities follows a repetitive pattern: read position, check relevance, understand requirements, assign product, mark special needs, note clarification questions. With 200+ positions per tender and multiple tenders per week, this process consumes significant capacity — precisely the personnel needed for costing and sales. The scaling problem is compounded by three factors: First, BoQ positions are not standardized — the same service can be described completely differently in different tenders. Second, requirements are scattered across different document sections — the BoQ contains the basic description, but technical details are in the building description, plans or appendices. Third, every position requires an expert decision: standard or custom? Deliverable or not? Risk acceptable or too high? These decisions can only be made by specialists — but the preparatory work of information extraction can be handled by AI.

How AI Extracts and Categorizes BoQ Positions

The AI-powered BoQ analysis process follows a structured sequence: 1. Document ingestion — GAEB files are natively parsed, PDFs are processed with layout recognition, scanned documents are digitized through OCR. The BoQ is fully digitized and structured. 2. Position extraction — every position is identified with position number, short text, long text, quantity, unit and hierarchical assignment (lot, trade, section). 3. Relevance filtering — based on the user's business profile, relevant positions are separated from irrelevant ones. A door manufacturer sees only door positions, a window manufacturer only window positions. 4. Requirement extraction — for each relevant position, the technical requirements are extracted: materials, dimensions, performance values, certifications, standards. 5. Product matching — the extracted requirements are compared against the user's product catalog: standard product, adaptation required, or custom construction. 6. Risk flagging — positions with incomplete specifications, contradictory requirements or exceptional risks are flagged. 7. Structured output — the result is a clear report with all relevant positions, their requirements and the product assignment.

Use Case: Door Manufacturer — Which Door Goes Where?

For a door manufacturer, a typical BoQ contains 200+ door-relevant positions: interior doors, exterior doors, fire doors (T30, T60, T90), smoke protection doors, security doors (RC2-RC6), sound insulation doors, multifunction doors and custom constructions such as tunnel doors or EBA-certified railway doors. The AI extracts every door position, identifies the required combination of functional requirements (fire + smoke + sound + burglar resistance), determines the installation situation (wall, frame, dimensions) and matches the result against the product catalog. Positions that combine multiple certifications are particularly highlighted — this is where the highest risk of miscalculation lies. The result: instead of spending hours working through the BoQ, sales receives in minutes a position list sorted by complexity with a clear standard/custom classification. Details on AI for door manufacturers at /en/blog/door-manufacturer-tender-ai.

Use Case: Precast — Standard Catalog vs. Custom Engineering

In precast concrete construction, the distinction between standard elements and custom constructions is the most important factor for costing. Standard elements — element walls, hollow-core slabs, standard columns — can be manufactured using existing molds. Custom constructions require new molds and individual reinforcement plans. The AI analyzes every position in the BoQ and categorizes: which elements are standard, which require modified standard molds, and which are genuine custom constructions? This three-tier classification gives engineering an immediate overview of the expected molding effort — the central cost driver in precast construction. Additionally, the AI identifies interface positions — connection details, joint designs, assembly sequences — which historically carry the greatest change order potential. Details on AI for precast manufacturers at /en/blog/precast-concrete-mold-calculation-ai.

Use Case: Windows — Variant Matching Across Dozens of Types

Window manufacturers face a particular combinatorial challenge: every position can require a unique mix of opening type (tilt-and-turn, sliding, fixed, top-hung), material (PVC, aluminum, wood, wood-aluminum), U-value, sound insulation, burglar resistance and design. With 100+ window positions per project, this creates a matrix that is nearly impossible to fully capture manually. The AI creates a requirement profile for each position and matches it against the product range. Positions whose requirement combination is not available in the standard range — such as a U-value of 0.8 W/(m²K) combined with RC3 classification — are immediately identified. This not only saves time but also prevents costing errors from overlooked requirements. The window-wall connection as the most common risk point is also included in the position analysis: where are the connection details specified, and where are they missing? Details on AI for window manufacturers at /en/blog/window-manufacturer-tender-ai.

Frequently Asked Questions