How BoQ automation actually works

  1. 01

    Ingest the BoQ — any format

    Excel files with any column layout, PDF cadernos de encargos, scanned bills of quantities, exported tables from CostX or RIB. The parser detects headers, units, item descriptions and quantities without manual template configuration. PaddleOCR handles scans; native text-extraction handles digital PDFs.

  2. 02

    Match every line against your history

    For each BoQ line, the engine queries your tenant ChromaDB of historical bids (hundreds of thousands of items, multilingual EN/PT/ES/FR). It returns the closest semantic + dictionary match plus a confidence score. A 300-line BoQ moves from upload to draft pricing in minutes — not the two-day spreadsheet ritual.

  3. 03

    Price every line — your suppliers first, then the open market

    The quote engine queries your private supplier price lists, then live retailer queries (Leroy Merlin, AKI, Würth, Wickes, Bauhaus etc. region-aware), then marketplaces. Returns ranked options by total cost delivered (unit + shipping + lead time + stock risk). Provenance preserved — every number has a "why".

  4. 04

    Validate, export, win

    Review each line with inline diff (AI draft vs your edit). Override with reason capture — that override feeds the model for next time. Export back to the client's preferred Excel template, to PDF for a bid submission, or to your ERP (PHC bridge available on Enterprise). Audit log captures every change with user, timestamp, IP — ISO 27001 A.12.4.1 conformant.

BoQ automation vs the alternatives

BoQ workflow is the European, Commonwealth and Indian construction standard. US-trained estimating tools force-fit it. Legacy EU tools (RIB CostX, Buildsoft, Causeway) handle it well but require manual cost-database upkeep and offer no AI matching. Here's where each category sits.

Excel + manual lookups

Where it shines: Free. Familiar. Total flexibility.

Where it hurts: Two days per BoQ. Knowledge lives in one estimator's head — leaves with them. No audit trail. Price database stale within weeks.

RIB CostX / Buildsoft / Causeway

Where it shines: Battle-tested. Strong drawing takeoff. CSI/NRM/SMM7-compliant.

Where it hurts: Five-figure perpetual licences. Steep training curve. No AI matching against your historical bids. No live retailer queries — you maintain the cost database manually.

Buildertrend / Procore Estimating

Where it shines: Integrated with full project-management suite.

Where it hurts: US workflow (CSI MasterFormat takeoff, not BoQ). EU contractors using BoQ workflow force-fit a square peg. Per-seat pricing scales painfully past 5 estimators.

Generic AI bidding tools (Togal, Beam, Handoff, CountBricks)

Where it shines: Modern UX. AI-first. Often free trial.

Where it hurts: US-trained models — coefficients drift on EU bids. No multilingual ingest. No private supplier lists. Pricing $35–$299/mo bottom tier; sub-€100/mo with EU billing rare.

What you get

Multi-format BoQ ingest

Excel (any layout), PDF (native or scanned), DXF/DWG drawings (block detection counts items automatically). Multi-format batch import — drop a folder, get one project.

AI matching trained on your bids

Per-tenant ChromaDB embedding store. Multilingual (EN/PT/ES/FR). Dictionary-based first (75% top-1), semantic fallback. The match gets sharper each project you close.

Private suppliers + live retailers

Your loaded price lists queried first. Region-aware live retailer queries (Würth, Leroy Merlin, AKI, Wickes, Screwfix, Bauhaus) for items not in your catalogue. Total cost delivered ranking.

Audit log + history ribbon

Every CUD operation captured with user, IP, microsecond timestamp, before/after — ISO 27001 A.12.4.1 conformant. Adobe-Fusion-style undo: revert any batch with one click.

Export how the client wants it

Excel preserving the customer's template, PDF bid-ready, CSV for downstream tools, JSON for ERP integrations (PHC bridge on Enterprise). Same BoQ in, same BoQ out — costed.

EU-hosted, GDPR-native

Servers in Portugal. AI inference local on owned GPUs. No third-party LLM API ever sees your BoQs. Signed DPA on Pro and above. SOC 2 on the roadmap for US-customer pipeline.

BoQ automation FAQ

What does BoQ automation actually mean?

Bill of Quantities (BoQ) automation is the replacement of manual line-by-line estimating work — identifying each item, finding the right historical price, looking up current supplier prices, formatting the output — with software that ingests the BoQ in its native format and produces a costed estimate in minutes. bidyou.ai automates ingest, AI-powered matching against your historical bids, multi-source pricing, and audit-grade export — leaving you in control of the final validate-and-ship step.

Does bidyou.ai support SMM7, NRM, CESMM4, or CSI MasterFormat?

The ingest engine is taxonomy-agnostic — it reads whatever your BoQ contains, in whatever order. For matching, you can map your taxonomy to ETIM or to your own coding scheme. SMM7 + NRM2 (UK), CESMM4 (civils), and CSI MasterFormat (US) are all supported via mapping. If your customer hands you a BoQ in their template, bidyou.ai matches against your history and exports back in the same template.

How does this differ from RIB CostX, Buildsoft Cubit, Causeway?

Legacy BoQ software gives you a powerful spreadsheet and a manually-maintained cost database. bidyou.ai gives you AI-powered matching against your own historical bids (the cost database builds itself from your closed projects) plus live retailer queries (your supplier prices update themselves). Same input, dramatically less manual upkeep. Where legacy software needs 1–2 weeks of training, bidyou.ai's UI is one-screen-per-project.

My BoQ is a scanned PDF. Can bidyou.ai still read it?

Yes. The ingest pipeline detects whether each PDF is native text or scanned image. Scanned BoQs go through PaddleOCR (multilingual OCR), then the structured extraction runs. Accuracy is high on cleanly-scanned modern PDFs; older photocopies may need a manual pass on edge cases.

Is there a free way to try BoQ automation before committing?

Yes — the Free tier is forever-free for 3 active projects, no credit card required. Drop in your first BoQ, see the matching against the seed database, and decide. Magic-link activation; you're in within a minute of signing up.

Where is my data stored and processed?

EU-based servers (Portugal) for Free / Starter / Pro tiers. All AI inference runs locally on owned GPUs — no third-party LLM API ever sees your BoQs. Enterprise tier supports on-premise deployment on your hardware. GDPR-native architecture; signed DPA on Pro and above.

Drop your first BoQ.
See it costed in minutes.

Free forever for 3 active projects. No credit card. Magic-link activation.

Sign up free See pricing