Recent workPilot

An AI roadmap and intelligence layer for a material passport platform

Agents that read supplier EPDs and BIM files, so passports take hours of review, not days of work.

Visual for an ai roadmap and intelligence layer for a material passport platformEurope's system of record for circular building materials
Problem statement

What the client was up against

Customers want the material passport to answer questions, not just document what exists. Behind the scenes, EPD ingestion, schema mapping and linking materials to BIM elements took 10 to 30 hours per building and capped growth.

Our approach

How we approached it

  1. Set an AI transformation roadmap with three outcomes
  2. Automated EPD extraction and validation against EN 15804+A2
  3. Proposed IFC to material links and routed only low confidence cases to people
  4. Designed an embedded intelligence tier customers can query
Solution we created

What we built

An AI-native 2.0 platform with an embedded intelligence layer. Agents read EPDs and IFC files, validate them, propose links, re check passports on a schedule and answer portfolio questions with cited evidence.

Agents in the system

The multi agent team behind it

SourcesClient platform, BIM and IFC files, EPD sources
EPD agentReads declarations and validates key fields
Linking agentMaps IFC elements to material families
Quality agentRe checks validity and completeness on a schedule
Intelligence agentAnswers portfolio, carbon and circularity questions
People decideReview, approve and act
EPD agentReads declarations and validates key fields
Linking agentMaps IFC elements to material families
Quality agentRe checks validity and completeness on a schedule
Intelligence agentAnswers portfolio, carbon and circularity questions
Client platformBIM and IFC filesEPD sources
What it was built to change

The outcome it is designed for

Linking effort modelled to fall from 10 to 30 hours to 1 to 3 review hours per building, continuously accurate passports and a new premium intelligence tier.

This is a pilot without published results yet. We show the problem and the build, not numbers we have not measured.

Client names, detailed case studies, demos and references can be shared on request.

Looking for a similar solution?

Tell us your version of this problem. We will walk you through what we built and what a first pilot would look like for you.

Let's talk