Experiment 02 / Trace blockers to their causes
Find what is holding the project back. Explore what can change.
Construction teams make decisions across design, procurement, fabrication, installation, and commissioning. Each decision can introduce dependencies that other participants may not see until their work is affected.
We’re exploring how AI agents can continuously examine those connections, identify emerging constraints, and prepare credible alternatives. Project professionals apply their judgment to the engineering, commercial, and delivery decisions. AI performs much of the investigation, comparison, coordination, and follow-through.
The questionCan AI distinguish necessary constraints from avoidable dependencies and help teams find a faster, less costly way to achieve the required outcome?
A principle across the lab
Valid and early signal.
Valid: establish the basis.
Use identifiable, current evidence relevant to the affected work. Keep confirmed facts, reported information, and AI inferences distinguishable. Identify missing or conflicting information.
Early: preserve time to act.
Identify a concern before a decision or commitment becomes difficult or costly to change, with enough time for the team to investigate and respond.
Prepare a useful response.
Explain what changed, the supporting evidence, what remains uncertain, the affected work, and when a response is needed to preserve the available options.
For this experiment, look for an emerging constraint before it stops work. Where the evidence is incomplete, agents prepare the specific inquiry or field check needed to establish the situation.
01 / The experiment
What we would test
Understand the work and its requirements.
Connect drawings, specifications, models, schedules, procurement records, field observations, and decisions to the affected systems, locations, and work packages. Preserve source references, revisions, and unresolved assumptions so the basis of each finding can be checked.
Watch for what needs attention.
When information changes, examine its implications across disciplines. Identify missing checks, conflicting requirements, unsupported assumptions, and approaching decisions before they interrupt work. Compare reported status with available field evidence, and identify where verification is still needed. Establish when the team would need to respond to protect the affected work, and prepare the next inquiry or action.
Question the dependency.
Establish why one activity must wait for another. Distinguish physical and technical requirements from contractual obligations, incomplete information, resource constraints, and established procedures. Examine whether the dependency can be removed, narrowed, or resolved differently.
Develop and test alternatives.
Explore changes to sequence, scope packaging, interfaces, component selection, and review arrangements. Evaluate which activities could proceed concurrently and what additional information would make that possible. Compare consequences for cost, completion, quality, safety, and rework using appropriate calculations and project evidence.
Prepare a coordinated decision.
Bring the affected disciplines together around a common analysis: the issue, supporting evidence, alternatives, recommendation, uncertainties, and judgments required. Prepare the proposed actions and record updates so an accepted decision can move directly into execution.
Check whether the work actually moves.
Track authorized actions and their results. Verify whether the constraint was removed, whether downstream work became ready, and whether the proposed improvement created problems elsewhere.
Agents keep checking the project as it changes, maintain the analysis, and bring people a prepared decision when their judgment is needed.
02 / An illustrative scenario
An equipment submittal is holding up related construction.
An air-handling unit submittal is awaiting approval. AI agents compare the outstanding review comments with the requirements for the roof curb, structural supports, electrical rough-in, controls, and commissioning to identify which work is actually affected.
Can some work be released while the remaining selections are resolved?
The agents check whether the manufacturer has confirmed the equipment dimensions, operating weight, support requirements, and service connection locations. They then prepare a proposal for the responsible design professionals and construction team to assess whether specific work, such as the roof curb and structural supports, can be released while controls selections remain open.
The proposal identifies the work that could proceed, the information and authorization required to release it, and the work that must remain on hold. It also explains what would need to be checked or reworked if the equipment selection changes.
03 / Evidence of value
What would establish value
The experiment would evaluate whether the analysis helps teams reach executable decisions and release productive work. Proposed benefits would need to be checked against actual outcomes.
- Supported concerns identified early enough for the team to take an effective action.
- A record of unsupported alerts, missed concerns, and the effort required to investigate findings.
- Less time spent assembling and reconciling information.
- Better-supported alternatives.
- Fewer unnecessary handoffs.
- Evidence that accepted decisions released productive work without transferring cost or risk downstream.