AI comes up in almost every business conversation now.
Companies feel like they need to be using it. Employees are worried about what it means for their jobs. Managers are concerned that competitors are moving faster.
In the graphic design industry, we hear a version of the same question all the time:
Why can’t AI just create the graphics?
The problem with that question is that it assumes the information provided at the beginning of a project is clean, accurate, and complete.
That is rarely the case in building automation.
The hard part is often not creating the graphic.
The hard part is figuring out what information is correct, what has changed, what is missing, and what the building actually looks like today.
The Information We Receive Is Not Always Clean
On some projects, we receive current Revit files, mechanical drawings, equipment submittals, point lists, and a well organized database.
On others, we may receive a phone picture of an old fire evacuation plan hanging on the wall.
That happens more than people would think, especially in older buildings being upgraded.
The original drawings may no longer exist. The building may have been remodeled several times. Rooms may have been added, removed, renamed, or repurposed without the original plans ever being updated.
A building owner may move departments, combine rooms, change room numbers, or turn an office into a lab or conference room.
The floor plan we receive may show one thing while the building is being used completely differently.
The reference material may include:
- Controls drawings
- Mechanical schedules
- Equipment submittals
- Architectural PDFs
- Revit or AutoCAD files
- Existing BAS graphics
- Database backups
- Point lists
- Screenshots
- Handwritten notes
- Technician markups
- Photos taken during a site visit
- Files created years apart by different contractors
Sometimes the only usable floor plan reference is a faded fire plan, a scanned blueprint, or a hand marked drawing.
That information still has to be reviewed, interpreted, and turned into something useful.
A picture of a fire plan may need to be corrected for perspective, redrawn, compared against mechanical drawings, and matched to current room names.
A submittal may show several equipment options without clearly identifying which one was installed.
A controls drawing may call an air handler one thing while the database calls it something else.
Before our team can create the graphic, someone has to decide what information can be trusted.
That is not just graphic design.
It is investigation, interpretation, and reconstruction.
The Database Keeps Changing
Even when we receive a database, it may not stay the same throughout the project.
Technicians make changes in the field as systems are installed, programmed, tested, and commissioned.
Points get renamed. Controllers are added or removed. Equipment assignments change. New objects are created. Old objects may remain in the database even though they are no longer being used.
Some of those changes happen directly in the live database without every drawing, point list, or project document being updated.
The building may also keep changing while the controls project is underway.
An owner may move people around, change room uses, add walls, remove walls, or renumber spaces.
The database may show part of the change. The floor plan may show another part. The controls drawings may still show the original design.
The newest file is not always the most accurate file.
That is where the idea of simply feeding everything into AI starts to fall apart.
When the information conflicts, something still has to determine which version is right.
That takes context and experience.
A Good Looking Graphic Can Still Be Wrong
AI can create impressive images and layouts.
That does not mean the finished result is technically correct.
A building automation graphic is not just a picture. It is an operator interface connected to live equipment, database points, commands, alarms, histories, schedules, and navigation.
A graphic can look complete and still have major problems.
- A room temperature may be connected to the wrong space.
- A command may be bound to the wrong piece of equipment.
- A renamed point may leave an object without a valid connection.
- A copied graphic may still contain references from another controller, floor, or building.
- Equipment may still appear on the screen even though it was removed from the final database.
The graphic may look right while giving the operator the wrong information.
That is why appearance alone does not tell you whether a BAS graphic is complete.
Creating the Graphic Is Only Part of the Job
The graphics may be built correctly from the information originally provided and still be wrong by the time the project is ready to go live.
That is why verification is so important.
The finished graphics need to be compared against the actual database to confirm the equipment, point mapping, bindings, labels, links, alarms, histories, and navigation are correct.
This part of the process is easy to overlook because the graphic already looks finished.
Visually, it may be done.
Technically, it may not be.
Verification is what confirms that the graphic matches the system that was actually installed, not just the information available when the project started.
This Is Where Good Tools Matter
QA Graphics is looking for every practical advantage that can make our work more efficient, consistent, and accurate.
We want AI and automation tools to keep advancing.
We also want to use the right tools for the right problems.
The answer is not to ask AI to guess how a building or database is configured.
The answer is to use purpose built tools to organize information, identify inconsistencies, reduce repetitive review, and help verify the final product.
QAGFoxhound was developed around that type of need.
It can support Niagara database validation, organization, point review, graphic reference checks, and system graphics verification.
That does not remove the need for an experienced person.
It gives that person a better way to review a large amount of information and focus on the areas most likely to create problems.
That is where technology creates real value.
Structure Has to Come Before Automation
The industry continues to talk about AI, digital twins, analytics, semantic tagging, and automated graphic generation.
All of those things depend on structured information.
Equipment needs to be classified consistently. Points need to be named or tagged in a way software can understand. Graphic assets need more information attached to them than just a file name or visual appearance.
That is where Vectortology™ fits.
Vectortology™ adds metadata and tagging to vector graphic assets so software does not have to guess what a symbol represents.
A pump should be identified as a pump. A fan should be identified as a fan. The asset should include enough structure for software to understand what it is and how it should be used.
This becomes more important as the industry moves toward Haystack aligned tagging, digital twins, automated workflows, and more intelligent software tools.
Vectortology™ does not fix incomplete drawings, field changes, or conflicting project information.
It helps create a better foundation once those assets are inside the BAS environment.
QAGFoxhound helps review and verify the database.
Vectortology™ helps bring structure and meaning to the graphic assets.
They solve different parts of the same larger problem.
We Are Not Against AI
QA Graphics is actively looking for ways to use AI and automation to improve our processes.
We expect the tools to get better, and we hope they do.
There are real opportunities to use technology to organize information, identify inconsistencies, reduce repetitive work, support database review, and improve verification.
But today, BAS graphic projects still require a lot of human interaction and decision making.
Someone still has to review the files, understand the mechanical system, identify conflicts, determine which information is reliable, ask questions, and verify that the finished graphics match the live database.
AI can help with parts of that process.
It cannot yet replace the judgment required to work through outdated drawings, incomplete databases, field changes, handwritten notes, pictures of fire plans, renamed points, and building layouts that may have changed several times.
We want every efficiency advantage we can get.
For now, the best results still come from combining useful technology with experienced people who understand what they are reviewing and why it matters.
The Better Question
The question should not be:
When will AI replace BAS graphic designers?
The better question is:
How can technology help experienced people work through messy project information, identify problems faster, and verify the final product?
That is where the real opportunity is.
In a specialized industry, the value is not just the ability to generate more content.
The value is knowing what is correct, what is outdated, what is missing, what changed, and what the final graphic needs to communicate.
That is the part of building automation graphics that AI FOMO continues to miss.
