Daily field briefing
POLR AIPOLR AI

AI + AEC Today

Daily AI intelligence for the built world

Weekly digest

Weekly AI in AEC stories

← Archive

Edition · Wednesday, July 22, 2026

AI & Construction

Weekly Pulse · Industry actionability

The field is piloting, not just experimenting.

60%
ExperimentalPilotingField-ready
  • AI now runs the whole first pass of a bid, not just parts of it
  • Back office AI is saving building firms dozens of hours a month
  • New AI is being built just for construction's messy files
  • Big AI models keep getting cheaper and faster

Practical Use of the Day

Turn a messy cost report into a clear update for each PM

Job cost reports are long and hard to read. AI can turn one into a short, plain update for each project manager. Paste in your cost export and ask for a per-job summary. Have it flag the jobs that are over budget and any costs that look coded wrong. Check its work, then send each PM just their part.

Try this prompt

You are helping a construction team read a job cost report. I will paste a cost export with many jobs and project managers. Do three things. First, write a short, plain summary for each project manager, covering only their jobs. Second, flag any job that is over its budget, and by how much. Third, list any costs that look coded to the wrong job or category so I can check them. Keep it simple and put each PM's summary under their name. Do not guess at missing numbers. Here is the report:

Tool — Claude (or any capable LLM)

Stories

4 curated
ConTechHigh impact

New AI does three weeks of bid work in an afternoon

Boon AI opened early use of a new tool called Boon Agent. It does the first pass of a bid on its own. It reads the specs, finds where drawings clash, and counts the work across trades. Then it sends out quote requests and hands your estimator a bid to check. On one real job, three weeks of work came back in about seven minutes.

Why it matters to GCs

Estimators are stretched thin, and each bid eats weeks. A tool that does the grunt work fast lets you bid more jobs with the same crew. Just make sure a person checks the numbers before they go out.

Boon AI·Jul 15·Read source →

LinkedIn
AI & BusinessModerate

Construction back office AI raises $15M to grow

Agave makes AI tools for the back office of building firms. It reads invoices, checks vendor papers, and pulls numbers from all your systems into one place. More than 500 contractors use it now. The company just raised $15 million to build more.

Why it matters to GCs

Back office work eats hours your team could spend on the job. Agave says users save more than 60 hours a month on data entry and reports. One firm turned a big cost report into a clear page for each of its 47 project managers in minutes.

Business Wire·Jul 7·Read source →

LinkedIn
ConTechModerate

Trunk Tools builds AI made just for messy job files

Most AI chat tools were built for clean text, not stacks of plans and specs. Trunk Tools took a different path. It cleans and links a job's files first, then trains its own AI on top. That lets its tools answer hard questions across millions of pages.

Why it matters to GCs

Field teams lose time hunting through drawings and specs for one answer. Trunk Tools says its setup cut some review jobs from about two months to ten days. Faster answers mean fewer delays and less rework.

TechBooky·Jul 6·Read source →

LinkedIn
AI FrontierModerate

Google ships a cheaper Gemini model and teases its next big one

Google put out three new Gemini AI models. The main one, Gemini 3.6 Flash, is faster and costs less than the last version. One of the others is built to find and fix security holes in code. Google also said it has started work on its next big model, Gemini 4.

Why it matters to GCs

Cheaper, faster AI means the tools your firm uses may cost less and run quicker. The tight race also keeps pushing prices down. It is smart to keep your options open and not lock into just one maker.

TechCrunch·Jul 21·Read source →

LinkedIn