What happens when AI works with construction administration software?

Artificial intelligence is becoming increasingly useful in architecture. AI assistants can summarize lengthy documents, compare product information against specifications, and help draft responses to project questions in seconds.

But most of these tasks involve working with information that someone has provided to the AI assistant. What happens when AI connects directly to the software architects use to manage their projects?

Imagine asking ChatGPT or Claude, “Which of our projects have the most overdue submittals?” or “Which consultants have the longest turnaround times?” and then getting an answer based on your firm’s actual project records—without having to export reports or compile information from multiple spreadsheets.

This is becoming possible through a technology called model context protocol (MCP), which allows AI assistants to connect to external software and access information stored there. For architecture firms, it’s an opportunity to make better use of the information they’re already collecting without changing how they manage their day-to-day work.

Construction administration software knows more than you might think

Consider a typical submittal. The document contains product information, technical details, and the architect’s review comments. An AI assistant can read that document and provide a useful summary. But the document alone doesn’t tell you when the submittal arrived, how long it spent with a consultant, how many times it was resubmitted, or whether the review exceeded its deadline. That information comes from the construction administration (CA) process itself.

Every time a submittal is received, an RFI is answered, or a field report is issued, additional information is generated about how the project is progressing. CA software captures much of this activity as part of the normal workflow. It tracks what’s outstanding, who’s responsible, when reviews are due, and what has happened along the way.

For the architects managing those projects, that information is essential to keeping work moving. But across dozens of projects, it can also reveal patterns that aren’t immediately obvious when looking at each project individually. Connecting AI to that information creates new possibilities for understanding what’s happening across a firm’s work.

Below are three ways AI and CA software can work together:

1. Creating a clearer picture across projects. For architects managing individual projects, CA software provides visibility into outstanding tasks, upcoming deadlines, and documents requiring attention. But what if you’re responsible for overseeing 20 or 30 projects at once? A principal or director of construction administration might want to know which projects have the most overdue submittals, where consultant reviews are taking longer than expected, or whether outstanding work is accumulating across particular projects.

An AI assistant connected to the firm’s CA data may help answer these questions without requiring someone to compile information from individual project reports. This gives firm leaders a broader view of project activity and helps them identify where additional attention or resources may be needed.

2. Finding opportunities to improve how teams work. Consider a firm that wants to understand why certain projects move through construction administration more efficiently than others. By comparing RFI and submittal turnaround times across projects, leadership can begin to identify differences in how teams work, how consultants respond, or how expectations are established. Similarly, reviewing recurring RFIs across projects may reveal specification details that could benefit from clarification.

These insights can help firms learn from their own project history, identify successful practices, and apply those lessons to future work. The goal isn’t simply to produce more reports. It’s to make better decisions using information the firm has already collected.

3. Reducing time spent preparing reports and updates. Many architecture firms still rely on manually maintained spreadsheets to track CA activity, particularly when leadership needs information across multiple projects. These reports can be valuable, but they take time to prepare and quickly become outdated.

When AI has access to current project data, it can help organize that information into summaries, reports, and meeting updates. For example, an architect preparing for an owner-architect-contractor meeting could ask an AI assistant to summarize outstanding RFIs and submittals, identify who’s responsible for the next action, and draft a meeting update. And a CA director could request a summary of overdue items across active projects without manually consolidating multiple project logs.

In these examples, the information comes from the CA system, while the AI assistant helps turn it into something useful.

AI doesn’t replace the need to manage the work

Connecting AI to CA software doesn’t eliminate the need for the software itself. Submittals still need to be received, assigned, reviewed, and issued. Deadlines need to be tracked. Consultants need to be coordinated. And firms need a reliable record of what happened, when it happened, and who took action. Those activities create the information that makes AI useful in the first place.

It’s also important that AI doesn’t take control away from the professionals responsible for the work. AI can help architects retrieve information, identify patterns, and prepare reports, while the CA system continues to manage the underlying workflows and project records.

As AI becomes more integrated into architectural practice, firms may find themselves using their software in new ways. Some tasks will continue to happen directly within specialized applications, while others, particularly reporting, analysis, and answering questions, may increasingly happen through AI assistants.

The opportunity isn’t to replace the tools architects rely on. It’s to make the information those tools capture more accessible and useful to everyone across the practice.

What if you could ask ChatGPT or Claude questions about your projects and get answers based on your actual CA data? Join Part3 for an upcoming webinar where we’ll demonstrate how AI assistants can connect to construction administration software, answer project questions, identify patterns, and help reduce manual reporting. We’ll also explain the technology making it possible, including MCP, with no technical background required. Register for the free webinar hosted by Part3.

AIA does not sponsor or endorse any enterprise, whether public or private, operated for profit. Further, no AIA officer, director, committee member, or employee, or any of its component organizations in his or her official capacity, is permitted to approve, sponsor, endorse, or do anything that may be deemed or construed to be an approval, sponsorship, or endorsement of any material of construction or any method or manner of handling, using, distributing, or dealing in any material or product.