Deep Compliance Review for a Major U.S. Research University

Finding what conventional design review misses

A major U.S. research university engaged Core Runway to evaluate how AI could improve compliance checking across complex construction documents.

The university maintains detailed design standards governing how buildings and infrastructure should be designed, specified and constructed. These requirements cover disciplines including mechanical, electrical, plumbing, architectural systems and equipment.

The challenge is familiar to large institutional owners: the standards are extensive, project documentation runs to thousands of pages, and determining whether every applicable requirement has been correctly reflected across drawings and specifications is extraordinarily difficult for human reviewers.

Core Runway was used to perform a deep compliance analysis of an existing project against the university's own design requirements.

The challenge

Traditional QA/QC relies heavily on experienced architects and engineers manually reviewing drawings, specifications and owner standards.

But the reviewer must simultaneously understand:

• which institutional rules apply to each part of the design;

• where the relevant information appears across hundreds of drawings and specification sections;

• whether a requirement has been explicitly satisfied;

• whether apparently compliant designs contain subtle contradictions;

• and whether different documents agree with one another.

General-purpose AI models face a similar problem. Simply loading thousands of pages into an LLM and asking it to "find compliance issues" produces relatively shallow results because the model has little understanding of which rules should be evaluated against which parts of the building.

Core Runway approaches the problem differently.

The Core Runway approach

The university's design standards were first converted into structured, traceable compliance requirements.

Core Runway's compliance ontology then associates those requirements with the relevant building systems, disciplines, equipment, materials and design information.

Project drawings and specifications are analyzed in the same structured way.

Instead of asking an AI model to reason across the entire document set at once, Core Runway can determine:

What should be checked, where it should be checked, and what evidence should demonstrate compliance.

The platform then performs thousands of targeted comparisons and returns findings linked to both the originating requirement and the relevant project documentation.

The result

The review identified more than 1,600 potential compliance issues across the project documentation.

These were not simply keyword mismatches or missing notes. Many required reasoning across different documents, drawings and engineering concepts.

Findings ranged from straightforward documentation omissions to potentially significant design conflicts requiring professional review.

Example findings

Missing chilled-water relief valves

University requirements called for relief valves on both the building and campus sides of chilled-water heat exchangers.

Core Runway located the relevant heat exchangers on the mechanical drawings and determined that the required relief valves were not indicated.

Emergency-power transformer requirement

The university standard required a dedicated dry-type, high-efficiency transformer for emergency power.

The electrical documents showed emergency transfer equipment and emergency panels, but the transformer schedule identified only a transformer associated with the normal-power system.

Core Runway flagged the absence of evidence for the required emergency transformer.

Prohibited pad-mounted transformer

The university's electrical requirements specified transformers located in vaults and prohibited installation on concrete pads.

Core Runway identified a drawing detail showing a transformer installed on a concrete pad and independently located the equipment schedule describing the same transformer as pad-mounted.

The system connected these pieces of evidence and flagged the configuration against the university requirement.

Non-compliant access doors

Mechanical requirements called for hinged access doors at specific dampers and reheat coils.

The architectural specification permitted lift-out gypsum access panels.

Although both documents were individually reasonable, Core Runway recognized that the permitted architectural product could conflict with the university's operational requirement for hinged access.

Incomplete floor-drain configuration

The university standard required several specific components within a floor-drain assembly.

Core Runway located the specified drain detail and identified multiple required elements that were not shown, including the required sediment-bucket configuration.

Incomplete air-terminal schedule

University standards specified the information that should be included in air-terminal schedules, including parameters such as room number, model, velocity, airflow, pressure drop and acoustic performance.

Core Runway compared that requirement with the project's schedule and identified required data fields that were absent.

Why this matters

No individual finding is the central story.

The important result is the scale and depth of review that becomes possible when institutional requirements, specifications and drawings can be evaluated systematically rather than sampled manually.

A human reviewer may know exactly what a compliant chilled-water system or electrical installation should look like. The difficulty is applying thousands of such requirements consistently across thousands of pages of project documentation.

Core Runway turns that institutional knowledge into a repeatable compliance process.

For the owner, that creates the opportunity to identify problems while they are still information problems—before they become RFIs, change orders, field rework or permanent departures from institutional standards.

From documents to institutional knowledge

The project also demonstrated an important difference between Core Runway and conventional document-search or general-purpose AI systems.

Core Runway does not treat an owner's design manual as simply another collection of PDFs.

The platform builds a structured representation of the relationships between:

owner requirements → building systems → specifications → drawings → equipment → materials → compliance evidence

That structure allows the AI to reason much more deeply about whether a design actually satisfies an owner's requirements.

It also creates a reusable institutional compliance layer that can be applied across future projects.

The opportunity

For major institutional owners managing hundreds of buildings and billions of dollars of capital investment, even very effective human QA/QC cannot examine every possible relationship between every requirement and every element of a design.

Core Runway is designed to augment those teams—not replace their professional judgment.

The platform gives architects, engineers and owners a much larger and more systematic set of issues to evaluate, while maintaining traceability back to the underlying source material.

The result is a fundamentally different level of design assurance: deep compliance review at a scale that was previously impractical.