Evidence Graph
Connect documents, revisions, measurements, assumptions, formulas, approvals and issued outputs.
Technology & Innovation
AI where interpretation helps. Deterministic engines where the number must be right.
CivilQuant is building a quantitative intelligence layer for the built environment: graphs, asset twins, semantic measurement, governed AI assistance and reproducible commercial calculations.
AI proposesExtraction, classification, matching and prioritisation
Engines calculateDecimal-safe, unit-aware quantitative arithmetic
Evidence provesSource, revision, rule, rate and checksum lineage
Reviewers authorizeExplicit gates around professional and monetary decisions
Core architecture
A connected quantitative model of the built asset lifecycle
CivilQuant is designed so that a drawing entity, measured quantity, BOQ item, asset component, inspection observation and capital intervention can belong to one connected evidence model.
Quantity Graph
Represent quantities as linked records with construction meaning, measurement provenance and downstream commercial relationships.
Asset Twin
Carry component identity, condition, cost, replacement and intervention evidence into the operating life of the asset.
Deterministic Formula Engine
Keep authoritative commercial arithmetic out of probabilistic language-model generation and inside restricted, reviewable calculation logic.
Semantic Measurement
Model walls, openings, rooms, slabs, beams, columns and other entities as construction objects rather than isolated marks.
Governance & Verify
Use evidence status, exceptions, approvals and reproducibility checks before information becomes commercial or professional truth.
AI-assisted, not AI-obscured
Use machine intelligence for interpretation and prioritisation
Built-environment information is messy: drawings, schedules, photographs, descriptions, rates and inspection notes rarely arrive as clean structured data. CivilQuant uses AI-assisted workflows where interpretation adds value while preserving a path back to the source.
- Drawing and document extraction candidates
- Semantic classification and matching
- Revision and change signals
- Anomaly and exception prioritisation
- Advisory forecasting and scenario support
- Human review before controlled outputs are issued
Deterministic quantitative core
Do not ask a language model to be the calculator of record
CivilQuant separates probabilistic interpretation from deterministic commercial arithmetic. Quantities, formulas, units, rates and issued calculations can be recomputed and reviewed from the underlying evidence.
- Decimal-safe arithmetic
- Unit-aware formulas
- Gross, deduction and net measurement logic
- Rate and assumption provenance
- Repeatable output checks
- Reviewer-controlled issue state
Innovation roadmap
Research directions are labelled as research, not sold as finished autonomy
CivilQuant can expand into stronger machine vision, graph intelligence, cost and condition forecasting, digital twins and automated exception triage as these capabilities reach the evidence and governance standard required for production use.
Visual intelligence
Research into richer understanding of drawings, site imagery and asset-condition evidence.
Graph intelligence
Use relationships between entities, quantities, assets and evidence to improve matching and anomaly detection.
Forecasting
Advisory models for cost, condition, exposure and capital scenarios with explicit uncertainty and reviewer control.
Digital asset twins
Keep component identity, evidence and lifecycle decisions connected across construction and operations.
Interoperability
API and integration patterns that let organisations connect CivilQuant to document, finance, BIM and asset systems.
Progressive autonomy
Increase automation only where exceptions, evidence and authorization remain visible and controllable.
Platform access
Technology you can actually use
CivilQuant is both a professional-services engine and a software platform. Create an account to use the workspace, or contact us to scope an enterprise or service engagement.