Inspiration
Inspiration
Construction projects generate large amounts of evidence: site photographs, inspection notes, drawings, technical documents, corrective actions, approvals and disputes.
In practice, this information is often fragmented across phones, email threads, messaging applications and separate reports. Important observations can lose their context, corrective actions may not be followed consistently, and it can become difficult to establish a reliable project history.
Systems Studio Evidence Node was conceived as a private, modular AI workspace for independent construction inspection and verifiable worksite evidence.
The project combines Systems Studio's experience in architectural design, MEP engineering and modular operational platforms with OpenAI technology.
What it does
Evidence Node is designed to transform worksite evidence into a structured inspection workflow.
A user can upload a construction photograph or enter an observation. The platform analyses the evidence and creates a structured inspection record containing:
- the relevant technical discipline;
- an objective observation;
- severity and confidence indicators;
- a proposed corrective action;
- required follow-up evidence;
- a recommended hold point where appropriate;
- a timestamped audit entry.
The resulting information is presented through a modular dashboard containing an inspection overview, evidence intake, issue register, hold points, an audit timeline and report generation.
The objective is not to replace the professional inspector. The AI assists with organisation, consistency, traceability and preparation of the inspection record, while final technical decisions remain under human control.
How we are building it
The project is being developed during OpenAI Build Week using Codex and GPT-5.6.
Codex is being used to help create and refine the application structure, frontend modules, backend endpoints, structured data contracts, testing and documentation.
GPT-5.6 is used to convert visual and written evidence into structured inspection information that can be reviewed by a human inspector.
The Build Week prototype is a focused vertical slice:
- upload or enter evidence;
- analyse the evidence;
- generate a structured issue;
- propose a corrective action or hold point;
- add the event to an audit timeline;
- generate an inspection summary.
The new Evidence Node is being built as a separate project using modular architectural patterns previously developed by Systems Studio. The inspection workflow, AI analysis pipeline and Evidence Node application are being created for this Build Week submission.
Challenges
The principal challenge is balancing useful AI assistance with professional responsibility.
Construction evidence can be incomplete or ambiguous. A photograph may not show the full installation, applicable specification or surrounding conditions. For this reason, the system must communicate uncertainty clearly, preserve the original evidence and keep the human inspector in control.
Other challenges include:
- producing consistent structured output from varied evidence;
- preventing unsupported technical conclusions;
- preserving traceability between evidence and generated observations;
- designing a workflow that is fast enough for use on a real worksite;
- maintaining privacy for NDA-sensitive projects.
What we learned
This section will be completed after the Build Week implementation and testing phase.
We are specifically evaluating how effectively Codex can accelerate development of a modular professional application and how GPT-5.6 can assist with evidence classification without obscuring uncertainty or human accountability.
What's next for Systems Studio Evidence Node
Following the Build Week prototype, the planned roadmap includes:
- client, contractor and independent-inspector roles;
- isolated project workspaces;
- private and local AI deployment options;
- immutable evidence records;
- mobile and offline evidence capture;
- drawing and specification references;
- change-order and hold-point workflows;
- versioned professional reports;
- managed subscriptions for architectural, engineering and construction projects.
About Systems Studio
Systems Studio is a Brussels-based engineering and AI systems company developing modular platforms for buildings, infrastructure, inspection and operational intelligence.
Evidence Node forms part of a wider Systems Studio ecosystem focused on private AI, intelligent infrastructure and professionally accountable workflows.
Log in or sign up for Devpost to join the conversation.