Gerome Inocencio monogramGEROME INOCENCIO
BUSINESS NOTES
Service Guides

AI Drawing Interpretation: Can Tools Like Trunk Tools Cortex Change Project Delivery?

By Gerome Inocencio10 min readBack to the blog ↗
AI Drawing Interpretation: Can Tools Like Trunk Tools Cortex Change Project Delivery? — Gerome Inocencio

AI Drawing Interpretation: Can Tools Like Trunk Tools Cortex Change Project Delivery?

AI-enabled construction management dashboard showing a digital project model, material tracking, and connected project data

The next advantage in construction may not come from buying another AI tool. It may come from making project information understandable, connected, and usable before decisions reach the field.

Construction projects generate enormous amounts of information: drawings, specifications, schedules, RFIs, submittals, change orders, procurement records, inspection reports, cost data, and site photographs.

Yet much of that information remains locked inside PDFs, disconnected platforms, email threads, and manually maintained spreadsheets.

This creates a practical question for contractors, developers, and project owners:

Can AI drawing interpretation turn static documents into a more reliable operating system for project delivery?

Tools such as Trunk Tools Cortex and BuildCheck suggest that the answer may be yes: but only when AI is treated as part of a governed workflow rather than as a replacement for professional judgment.

The real problem is not the drawing

A drawing is not merely an image. It contains decisions.

A single plan set may communicate:

  • Dimensions and locations
  • Materials and assemblies
  • Door, window, equipment, and fixture information
  • Discipline coordination requirements
  • Revision history
  • Detail references
  • Notes and specifications
  • Relationships between architectural, structural, civil, and MEP systems

Traditionally, project teams extract this information manually. Estimators review sheets for quantities. Engineers check coordination. Project managers search for revisions. Site teams interpret details while work is already moving.

That process can work, but it creates several risks:

  1. Important information may be missed.
  2. Different teams may work from different revisions.
  3. A change on one sheet may not be connected to its impact elsewhere.
  4. Field decisions may rely on incomplete context.
  5. Errors may be discovered only after they create RFIs, delays, rework, or cost exposure.

AI drawing interpretation aims to address this by converting visual and textual information into structured, searchable project data.

What AI drawing interpretation actually does

At a basic level, drawing-interpretation systems use technologies such as optical character recognition, computer vision, document classification, and multimodal analysis.

The goal is to identify objects, labels, dimensions, references, revisions, and relationships within a drawing set.

The result is not simply a better PDF search. The more important outcome is the creation of structured project information that can potentially connect with estimating, scheduling, procurement, BIM, quality assurance, and field-management workflows.

For example, a system may help identify:

  • Where a specific door type appears
  • Which sheets reference a particular detail
  • Whether a revised drawing changes a room or assembly
  • Which specifications relate to a selected object
  • Whether a note, dimension, or callout is missing or inconsistent
  • Which project records may be affected by a design change

However, the quality of the result depends on the quality of the source documents, the system’s capabilities, and the review process established by the project team.

What ENR’s research brief tells contractors

An Engineering News-Record research brief makes a central point: contractors seeking practical value from AI must first organize their data.

The article attributes this view to Brett Poulos, national director of preconstruction and estimating at Burns & McDonnell. According to ENR, the firm emphasizes structured data, clear governance, employee training, and pilot programs before broad deployment.

The brief also reports on a manufacturing expansion project where Burns & McDonnell used reality capture, continuously updated models, and AI-enabled progress tracking. ENR reported that this approach helped identify issues earlier, plan shutdowns more effectively, and reduce rework before problems reached the field.

The broader lesson is important:

AI does not remove the need for project controls. It increases the value of project controls.

If drawings, schedules, costs, and field records remain disconnected, AI may only produce faster answers from incomplete information. If those records are structured and governed, AI can help teams identify relationships and risks earlier.

Construction professionals collaborating around project information and workflow planning

Where Trunk Tools Cortex may fit

Trunk Tools describes Cortex as an AI layer for construction information. Its published materials position the platform as capable of interpreting drawings and connecting project records across documents such as specifications, RFIs, submittals, schedules, and change orders.

The proposed value is not limited to asking questions about a PDF.

A connected document-intelligence layer could allow a team to:

  • Select an object on a drawing and review related project information
  • Compare drawing revisions
  • Surface changes that may not be enclosed by revision clouds
  • Search across multiple document types
  • Connect drawing elements with RFIs, submittals, schedules, and specifications
  • Support bid, submittal, and RFI workflows

These capabilities could change project delivery by moving teams from document retrieval to context-based decision-making.

For example, instead of asking, “Where is this detail located?” a project manager may eventually ask:

  • What changed between the previous and current revision?
  • Which procurement packages could be affected?
  • Are there open RFIs connected to this location?
  • Which scheduled activities depend on the revised detail?
  • Has the field already installed work based on the earlier information?

NEEDS_SOURCE_CHECK: The capabilities above are based on Trunk Tools’ published product descriptions and should be independently validated through product documentation, a controlled demonstration, customer references, or project-specific testing.

Where BuildCheck may fit

BuildCheck positions its platform as an AI-powered drawing review system for architectural, structural, MEP, civil, and related drawing sets.

Its official materials describe automated checks for drawing health, coordination issues, omissions, and logical inconsistencies. The platform also describes results being pinned directly to drawings so teams can review and track potential issues.

This type of tool may be most useful before construction begins or before a design package is released for pricing, permitting, procurement, or execution.

Potential applications include:

  • Identifying missing dimensions or tags
  • Checking detail references
  • Surfacing certain 2D coordination conflicts
  • Flagging possible omissions
  • Reviewing consistency between drawings
  • Creating a structured list of items for design coordination

The practical benefit is not that an AI system becomes the approving engineer or architect. The benefit is that it may provide an additional review layer that helps professionals focus their attention on higher-risk issues.

NEEDS_SOURCE_CHECK: BuildCheck’s stated number of automated checks, discipline coverage, accuracy, and impact on RFIs or change orders require independent confirmation. Vendor descriptions should be treated as product claims until supported by third-party testing or project evidence.

Digital project-management dashboard representing project oversight, document control, and quality review

How this could change project delivery

1. Earlier issue detection

When drawing information can be searched and compared systematically, teams may find inconsistencies before they become field problems.

Earlier detection can create more options. A design issue discovered during preconstruction may be easier to address than one discovered after procurement, fabrication, installation, or commissioning.

2. Better revision control

A project’s source of truth should not depend on which file was last downloaded or which email attachment a person happened to open.

Drawing intelligence can support a more visible revision process by connecting current sheets with previous versions, affected locations, and related project records.

Human review remains essential, but the system can help direct attention to what changed.

3. More useful RFIs

A well-prepared RFI requires context. It should identify the location, relevant drawing or specification, observed conflict, and decision required.

If an AI system can search connected project documents, it may help teams prepare more complete questions and avoid submitting RFIs that have already been answered elsewhere.

NEEDS_SOURCE_CHECK: Whether a specific platform consistently reduces unnecessary RFIs or improves response times must be measured on actual projects.

4. Stronger procurement coordination

Drawing changes can affect quantities, equipment, materials, and long-lead items.

Connecting drawing objects to procurement records may help project teams ask better questions earlier:

  • Has the item been ordered?
  • Has fabrication started?
  • Does the selected product match the latest requirement?
  • Is a substitution required?
  • Will the change affect the schedule or budget?

This is where drawing interpretation becomes more than a design tool. It becomes part of commercial and operational control.

5. More visible owner reporting

Owners do not need more disconnected reports. They need a clearer view of what has changed, what is at risk, and which decisions require attention.

A connected project-data layer could support reporting around:

  • Open design decisions
  • Affected work areas
  • Procurement exposure
  • Schedule dependencies
  • Cost implications
  • Outstanding approvals
  • Verified versus unverified information

That visibility is valuable only when the underlying data is traceable.

Aerial view of a large-scale industrial construction project with multiple work fronts and structures

AI should recommend; people should approve

The most important control is not the tool. It is the approval architecture around the tool.

A responsible workflow should define:

  • Which documents may be uploaded
  • Who can access project information
  • How revisions are identified
  • Which AI outputs require engineering or management review
  • How false positives and missed issues are recorded
  • How decisions are documented
  • What becomes part of the official project record

AI can highlight a possible conflict. It should not quietly authorize a design change.

AI can summarize a revision. It should not replace the responsible professional’s review.

AI can identify a procurement risk. It should not commit the company to a purchase or client decision without approval.

This is the difference between using AI as a controlled operating layer and using it as an ungoverned shortcut.

A practical pilot for contractors and owners

A sensible pilot does not begin with every project document.

Start with one bounded workflow, such as:

  1. Select a representative drawing package.
  2. Include clean PDFs, scanned sheets, multiple disciplines, and at least one revision set.
  3. Define the specific fields or issues the tool must identify.
  4. Compare AI outputs with a human-reviewed baseline.
  5. Record false positives, missed issues, and review time.
  6. Test whether outputs connect to existing project controls.
  7. Establish access, privacy, and approval rules.
  8. Decide whether the result is useful enough to expand.

This approach follows the discipline described in ENR’s brief: organize the data, train the people, establish governance, and test the application before scaling it.

At IGRIS Balteus Construction, our flexi-systemic approach is built around adapting the delivery method to the project, client, budget, and risk profile. That same principle applies to AI adoption. The right question is not, “How do we use AI everywhere?”

The better question is:

Which decision can we make earlier, with better evidence, if our project information is connected?

Source-check section

The following claims require verification before publication or formal use:

  • Trunk Tools Cortex can interpret approximately 400 construction object types.
  • Cortex detects drawing changes outside revision clouds.
  • Cortex connects drawings, specifications, RFIs, submittals, schedules, and change orders through a knowledge graph.
  • BuildCheck performs 300 or more automated checks.
  • BuildCheck’s checks cover all stated disciplines and drawing types.
  • Either platform measurably reduces RFIs, change orders, rework, or project duration.
  • AI drawing interpretation can reliably feed estimating, scheduling, procurement, BIM, or QA/QC without project-specific configuration.
  • Any accuracy, return-on-investment, or productivity figures supplied by vendors.

Primary reference lanes:

ChatGPT Work banner prompt

Create a premium editorial LinkedIn banner about AI drawing interpretation in construction. Show a realistic project-control workspace where a construction professional reviews a large-format architectural and MEP drawing set beside a connected document-intelligence dashboard. The dashboard should show a drawing revision comparison, highlighted plan locations, linked specifications, an RFI panel, procurement status, and a human approval queue. Use authentic commercial photography with natural office lighting, realistic monitors, paper drawings, construction documents, and professional safety-conscious attire. Do not show robots, cyberpunk effects, floating holograms, generic AI stock imagery, or fabricated software logos. Use IGRIS brand colors: primary navy #0D2040, deep navy #002147, teal #0D9488, copper #B87333, light page #EDF2F8, white #FFFFFF, and light border #E2E8F0. Use Montserrat for interface labels and Lora sparingly for an editorial emphasis. Leave clean negative space on the left for the headline: “AI Can Read the Drawing. Who Approves the Decision?” Format 16:9.

Scale your business with AI

Are you ready to scale? Is your Enterprise AI-ready?

Explore Gerome Inocencio’s work and ideas at https://geromeinocencio.figma.site/.

If your organization wants to move from disconnected documents and reactive decisions toward connected systems, source-of-truth reporting, and practical human-approved AI workflows, start with one process, one dataset, and one measurable outcome.

Scale your business with AI. Scale your business with me.

Gerome Inocencio logo

About the author

Gerome Inocencio

Gerome writes from the intersection of construction, markets, and accountable delivery in the Philippines. His work is grounded in the questions owners and project teams need answered before the next commitment.

Keep reading

More from this desk.

All posts ↗
AI in Construction: What Owners Should Adopt First — Gerome Inocencio

Service Guides

AI in Construction: What Owners Should Adopt First

This business note from Gerome Inocencio looks at AI in Construction: What Owners Should Adopt First through a practical project lens. Use the guidance to clarify scope, cost, procurement, quality, and the next decision before work begins.

Business notes

Useful writing for the next decision.

Occasional notes on construction, markets, materials, and doing business in the Philippines. No noise, no inflated promises.

Talk through the next move

Bring the project into focus.

Clear costing, no hidden charges.