DBIA 2026 Day 3: What Booth Conversations Taught Us About Construction AI

Another round of interesting discussions took place at Booth 131 on the last day of DBIA 2026. Design-build experts continued exploring iFieldSmart AI, discussing real-world applications of AI and sharing project challenges.

Day 3 offered new insights into AI’s potential applications in construction, from integrating data across project papers to automating tedious processes and putting discoveries into practice.

We examined the discussions and questions that were particularly noteworthy on the conference’s final day. These are some of the main themes we heard, along with their implications for the changing requirements of design-build teams.

From AI Curiosity to Real-World Application

At most tech events, the first thing people say is “show me a demo.” This time visitors went straight to handing over real assignments. Build the logs. Check documents against requirements. Draft the reports. Log a field issue and send it to the right trade.

That’s a different kind of conversation. Curiosity gave way to delegation. Builders already know where their hours go, so they went directly to those tasks. It also puts pressure on anyone building these tools, because a vague answer is no use to somebody with a deadline.

Different Roles, Shared Project Challenges

A strange thing: the same underlying mess looks different depending on where you sit.

A general contractor wants to know if a submittal complies. A subcontractor wonders whether open RFIs will slow the crew down. An architect needs to know which documents a design change touches. An owner wants a look at the jobsite and a sense of what to do next.

Four chairs, one tangle. Documents, models, and field conditions that don’t agree with each other. It’s probably why single-purpose tools disappoint so many teams. The value lives in the links, like RFI to spec, spec to submittal, field to model.

Why Comparing Project Information Matters

Reading a document with AI is the solved part. What caught our eye were requests to compare. Project specs against an owner’s own standards. Submittals against their spec sections. Drawing sets from different dates. Design changes against open RFIs and submittals.

Comparison is where hours disappear, and mistakes hide, since the differences are tiny and buried. We expect a lot of useful construction AI to look like this: a careful second pair of eyes pointing at the few spots that deserve a human’s attention.

Turning AI Recommendations into Action

Count how many requests came in two parts. A review, then a draft response. A list of overdue items, then a nudge to the people responsible. A field issue, then an assignment to the right trade. A meeting, then a list of who owes what.

The second half is where a project actually moves. A finding stuck in a chat window helps nobody. The tools that earn a spot on a jobsite will carry a finding into the next step, with a person signing off before anything leaves the building. That’s the heart of agentic AI in construction. New to the term? Our glossary explains the AI agent.

Connecting the Jobsite with the Digital Model

AI’s potential to give teams a better understanding of what’s happening on site in comparison to what’s depicted in the design was another topic that came up during the discussions.

Examining recent jobsite photos, comparing installed work to BIM models, spotting inconsistencies, and confirming coordination details like installed elevations piqued visitors’ interest. Additionally, they wanted findings to be connected to particular locations so that teams could promptly identify problem areas and areas that required attention.

This means that it is becoming increasingly important to combine model and field data. Construction teams want to connect both types of data rather than treat one independently from the other to make better decisions.

The reliability of these comparisons relies heavily on the quality of data used for them. Models that do not pass the test of time or are not created in full can cause distortions in the outcomes. Hence, while AI devices can connect all the information, it can still be quite important to follow the status of the project.

Workflow Ideas from the Design-Build Community

Some of our favorite moments were requests for things we haven’t built. Revision history tracking. Integrated MEP coordination. Change order tracking that separates out-of-scope work from what a team absorbed. Spec compliance against an owner’s standards. Quantities pulled from the model. Site access control. One owner even asked about a post-construction performance workflow.

Look closer, and they share a pattern. Each starts with information that already exists in different places and asks for it to be put to work. It’s a good hint that useful construction AI has to bend to the way each firm runs its jobs, which is the reason we offer Custom AI Workflows.

Putting Connected Construction AI into Practice

In iFieldSmart AI, RFIs, drawings, specifications, project management platforms and BIM models are all incorporated into a single AI-based cycle. It interacts seamlessly with the programs and tools being used by your teams, such as Procore, Box, SharePoint, Google Drive, and others.

Constructability Review, Scope Gap Analysis, Submittal Compliance Review, and Change Management AI are some of the notable features of AI. It allows teams to review the findings, make decisions, and stay in control of every action.

Taking the Conversation Beyond DBIA 2026

As DBIA 2026 comes to a close, we’re taking away more than just three days of conversations. We’re leaving with new ideas, valuable feedback, and a better understanding of the challenges design-build teams want AI to help solve.

You can discover what we’re working on. Explore our AI construction skills, find answers in our construction AI, or visit our DBIA 2026 event page.

The event may have wrapped up, but the conversations around making construction smarter and more connected are only getting started.