Compliance Does Not End When the Design Is Finished

A construction project can comply perfectly with an owner's standards on paper and still end up with the wrong product installed in the building.

Why?

Because design compliance and construction compliance are often treated as separate processes.

They shouldn't be.

The Requirement Starts Upstream

Consider a relatively simple product such as paint.

The owner's standards may establish acceptable manufacturers, sustainability requirements or performance criteria.

The project specification then defines the required coating system.

The drawings may identify locations, colors and finishes.

The subcontractor prepares a submittal.

The manufacturer supplies product data, technical documentation and environmental information.

The architect or engineer reviews the package.

Eventually, someone installs the product.

The important point is that the requirement has traveled through multiple documents and multiple organizations before becoming something physical.

At every handoff there is an opportunity for information to be lost.

Specifications Are Only One Link in the Chain

Traditional document review tends to look at each stage separately.

The specification is checked during design.

Submittals are reviewed during construction.

Sustainability documentation may be reviewed by another consultant.

Owner requirements may be checked by the facilities team.

Each reviewer is looking at a different piece of the same decision.

But the real question is simpler:

Does the thing we are actually going to install satisfy what the owner originally required?

Answering that question requires connecting the documents rather than reviewing them in isolation.

The Submittal Is Where Theory Becomes Reality

During design, a specification might require:

• a particular performance level

• an approved manufacturer

• a material composition

• a VOC threshold

• a finish

• a warranty

• specific testing

• particular supporting documentation

But the specification is still an instruction.

The submittal introduces the actual proposed product.

That makes the submittal process one of the most important compliance checkpoints in construction.

It is where:

“This is what the project requires”

meets:

“This is what we intend to buy and install.”

Yet much of that comparison is still performed manually.

The Information Is Spread Everywhere

A reviewer may need to move between:

• owner design standards

• project specifications

• drawings and schedules

• approved manufacturer lists

• product data sheets

• environmental documentation

• shop drawings

• previous review comments

Even apparently straightforward scopes can involve dozens of individual requirements.

The difficulty is not necessarily that any single requirement is hard to understand.

The difficulty is finding all of them reliably.

Moving the Review Earlier

There is an opportunity to improve the process before the submittal even exists.

Instead of waiting for a subcontractor or manufacturer to assemble a package and then discovering what is missing, AI can first extract the applicable requirements from the project documents.

Those requirements can become a structured pre-submittal checklist.

For example:

Required product → location → performance criteria → approved manufacturers → documentation required

The subcontractor or manufacturer begins with a clear definition of what the project expects.

Then, when the actual submittal package comes back, it can be compared automatically against those same requirements.

The workflow becomes:

Requirements → submittal preparation → manufacturer evidence → automated comparison → human review

That is fundamentally different from simply using AI to summarize a product data sheet.

Connect the Owner All the Way to the Product

For institutional building owners, the bigger opportunity is connecting this workflow all the way upstream.

Imagine a single chain:

Owner standards

↓

Project specifications and drawings

↓

Submittal requirements

↓

Proposed manufacturer and product

↓

Supporting evidence

↓

Human approval

Every decision remains traceable to the requirement that created it.

Now a facilities team can ask not merely whether a project specification complied with its standards during design, but whether the actual products proposed during construction still comply.

Compliance Should Be Continuous

Construction documentation is constantly changing.

So compliance cannot realistically be treated as a one-time event.

A project may be compliant when the specification is issued and become non-compliant after a substitution.

A compliant design may generate a non-compliant submittal.

A product that meets the technical specification may conflict with an institutional preference that was never carried forward correctly.

The solution is not simply another review step.

It is maintaining the connection between requirements and decisions throughout the project.

AI makes that increasingly practical.

The goal is not to remove architects, engineers, contractors or facilities professionals from the approval process.

It is to give them something they have rarely had before:

A continuous, source-linked line from the owner's rulebook to what eventually goes into the building.

Your Design Standards Are Only as Good as the Last Project That Followed Them

Universities, healthcare systems and other large building owners often maintain extensive internal design standards.

They exist for good reasons.

An institution may have decades of experience operating hundreds of buildings. It knows which equipment performs reliably, which materials cause maintenance problems, which systems its facilities teams can support, and which design decisions create unnecessary operating costs.

That knowledge gets turned into design guidelines, technical standards, preferred manufacturers, standard details and specifications.

The problem is not creating the rulebook.

The problem is making sure every project actually follows it.

The Scale Problem

Consider what an architectural or engineering team may be asked to reconcile.

On one side is the owner's design manual — potentially hundreds or thousands of requirements covering architectural systems, mechanical equipment, electrical infrastructure, plumbing, controls, finishes, sustainability requirements and preferred products.

On the other is a project specification that can itself run to several thousand pages, accompanied by hundreds of drawings.

Every applicable requirement has to make it across that gap.

And the comparison is not simple keyword matching.

A standard might require a particular material, prohibit another, specify a preferred manufacturer, establish a performance threshold, require documentation, or allow an exception under certain circumstances.

Someone has to understand both documents well enough to identify the relationship.

Standards Change. Projects Change.

The problem becomes harder because neither side is static.

Institutional standards evolve as facilities teams learn from completed buildings, products change, regulations are revised and operating priorities develop.

Project documentation changes too.

Specifications are revised. Addenda are issued. Products are substituted. Value engineering changes decisions. Submittals introduce the actual products that will eventually be installed.

That means compliance is not a single review performed at one point in the project.

It is a continuing comparison between what the institution requires and what the project is proposing.

Why Previous Projects Are Not a Reliable Compliance System

One common approach is to begin a new project using specifications from a previous project.

That is understandable. Reusing proven work is efficient.

But it creates a hidden risk.

The previous specification may have been based on an older version of the owner's standards. It may contain an approved exception that applied only to that project. It may contain something that was never compliant in the first place.

Once copied, those decisions can propagate from project to project.

The specification begins functioning as an informal version of the owner's rulebook.

The institution now has two standards:

The one it publishes.

And:

The one its projects have historically used.

Those two can slowly diverge.

The Institutional Knowledge Problem

There is another complication.

The written rulebook is rarely the complete rulebook.

Experienced facilities personnel know things that have never been formally documented.

A particular product may technically comply but have performed poorly in the field.

A specified system may require an exception in a particular building type.

A facilities team may prefer one solution because its technicians already stock parts and know how to maintain it.

These decisions often live in people's memories, emails and previous project files.

When experienced employees retire or leave, some of that knowledge leaves with them.

The next team has to rediscover it.

AI Changes What Is Practical

Until recently, comparing every requirement against every relevant section of a multi-thousand-page project package was technically possible but economically difficult.

The labor required made comprehensive review impractical.

AI changes that equation.

A system can now ingest an institutional rulebook, analyze a complete project specification and systematically identify where requirements appear to be satisfied, contradicted or only partially addressed.

Importantly, AI should not make the final professional decision.

Its job is to perform the exhaustive research.

A qualified human reviewer remains responsible for determining whether a design is acceptable.

The useful model is therefore:

AI finds and compares. Humans decide.

From Static Rulebook to Living Institutional Knowledge

The bigger opportunity goes beyond faster document review.

Imagine that every time a reviewer approves an exception, rejects a product, clarifies a requirement or establishes a preferred approach, that decision can become reusable institutional knowledge.

The next project starts with what the organization learned on the previous one.

Instead of repeatedly interpreting the same rules, institutional knowledge accumulates.

The rulebook becomes more than a PDF.

It becomes a living system connecting:

Standards → project designs → products → evidence → human decisions → future standards

That may ultimately be the most important application of AI in institutional construction.

Not replacing professional judgment.

Making sure that professional judgment is applied consistently — and that what the organization learns is never lost.

Beyond Data Extraction: Why Multi-Variant EPDs Require Real Intelligence

One of the biggest misconceptions in sustainability compliance is that Environmental Product Declarations (EPDs) are simple documents.

They are not.

While extracting data from a single-product EPD can often be straightforward, many manufacturers publish multi-variant EPDs that cover dozens—or even hundreds—of product configurations within a single document. In these cases, identifying the correct sustainability values requires far more than finding a number on a page.

The Hidden Complexity of Multi-Variant EPDs

Consider a roofing manufacturer that publishes one EPD covering both standard roofing tiles and tapered roofing tiles.

At first glance, you might assume the recycled content for the tapered tile is listed directly in the document.

In reality, it often isn't.

Instead, the EPD may provide:

• A table of product dimensions

• Thickness ranges for various tile types

• Average thickness calculations

• Environmental performance data for standard products only

• Instructions explaining how variant products relate back to the standard product data

To determine the recycled content of a tapered tile, you may need to:

1. Identify the tapered tile product.

2. Calculate its average thickness based on the dimensions provided.

3. Map that average thickness to the corresponding standard tile category.

4. Locate the recycled content value associated with that standard product.

5. Apply the correct result to the tapered tile.

This is not document extraction.

It is document interpretation.

Why Traditional Software Struggles

Most sustainability platforms and document processing tools are designed to extract values that already exist in a structured format.

They can find:

• Recycled content percentages

• Global Warming Potential (GWP)

• VOC levels

• Product names

What they cannot reliably do is follow the reasoning process required when the answer must be derived through calculations, cross-references, and interpretation of manufacturer guidance.

As a result, sustainability consultants and project teams often perform these calculations manually, introducing delays and increasing the risk of errors.

How ConstructIQ Solves the Problem

We recently expanded ConstructIQ's sustainability intelligence engine to support multi-variant EPD interpretation.

Instead of simply extracting values from a document, ConstructIQ can now:

• Identify product variants covered by a shared EPD

• Understand manufacturer-specific calculation methodologies

• Perform intermediate calculations automatically

• Cross-reference multiple tables and sections

• Determine the correct sustainability values for the exact product selected

In the tapered roofing tile example, ConstructIQ automatically calculates the average thickness, identifies the matching standard tile category, and returns the correct recycled content value—without requiring manual intervention.

From Document Processing to Sustainability Intelligence

This capability highlights an important distinction.

The future of sustainability compliance is not about building better OCR systems or faster document extraction tools.

The real challenge is transforming complex manufacturer documentation into usable project intelligence.

Every industry has products where sustainability data is hidden behind formulas, lookup tables, conditional logic, or engineering assumptions. Solving these challenges requires systems that can reason through documents the same way an experienced sustainability consultant would.

That is the direction we are building toward.

Because finding information is useful.

Understanding how to derive the right information is where the real value begins.