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, August 26, 2026

AI & Construction

Weekly Pulse · Industry actionability

The field is piloting, not just experimenting.

55%
ExperimentalPilotingField-ready
  • Big firms start to trust AI with drawing and code checks
  • AI now spots when the jobsite does not match the plan
  • Cheaper, faster AI chips push more AI into your tools

Practical Use of the Day

Turn messy field notes into a clean daily log

At the end of the day your notes are short and messy. Copy them into an AI tool. Ask it to write a clear daily log. Tell it to list work done, crews on site, weather, delays, and any safety issues. You get a neat log in seconds. Read it against your notes before you save it, so nothing is added or missed.

Try this prompt

I am a general contractor. Turn my rough field notes into a clear daily log. List work done, crews and head counts, weather, delays, deliveries, and any safety issues. Use only what is in my notes. Keep it short and plain. Flag anything that looks missing or unclear. Here are my notes: [paste them].

Tool — Claude (or any capable LLM)

Stories

4 curated
ConTechHigh impact

Arcadis backs Nomic, an AI built to do design and construction paperwork

Nomic is an AI tool made just for design and construction firms. It uses AI helpers called agents to do slow office work. They check drawings against a firm's own rules. They look for code problems. They review submittals and dig up answers for RFIs. A big engineering firm named Arcadis tested it for six months in 12 countries with about 150 people. Most said it changed how they work. Now Arcadis has put money into the company.

Why it matters to GCs

This is the kind of paperwork that eats hours on every job. If AI can do a first pass on drawing checks and RFIs, your team gets time back and catches issues sooner. When a firm this size buys in, it is a sign the tools are getting good enough to trust on real work.

AEC Magazine·Aug 3·Read source →

LinkedIn
ConTechModerate

Buildots adds laser scans so its AI can catch site mistakes faster

Buildots uses AI to track how a job is going. Workers walk the site with a 360 camera. The AI turns those photos into a clear record of what is built. Now Buildots has teamed up with a company called NavVis to add laser scans. The scans are very exact. The AI can now spot when the real building does not match the plan, and it does it faster. It is already in use on an Intel chip plant job.

Why it matters to GCs

Catching a wrong install early is the difference between a quick fix and a costly tear-out. This tool watches the site and flags clashes and off-size work before they pile up. On tight jobs like data centers and hospitals, that can save weeks and real money.

AEC Magazine·Aug 11·Read source →

LinkedIn
AI FrontierModerate

Claude's AI helper now keeps one memory across chat and work

Anthropic makes an AI helper called Claude. It now keeps one memory across both its chat and its work mode. Before, you had to tell it your project details twice when you switched. Now it remembers things like a job's name, key dates, and people. So you can plan in chat and then send it off to do the task without starting over. It leaves out private things like health or ID numbers unless you say it is okay.

Why it matters to GCs

Telling an AI the same job facts over and over wastes time and leads to slips. A tool that keeps your project straight is faster and safer to lean on for daily paperwork. Just check what it stores before you feed it client details.

TechCrunch·Aug 25·Read source →

LinkedIn
AI FrontierModerate

OpenAI's new Jalapeno chip is built to run AI fast and cheap

OpenAI built a new computer chip called Jalapeno with help from Broadcom. The chip is made to run AI fast and cheap. Early tests show it can do more AI work per unit of power than top chips from Nvidia. It also gives answers more quickly. It starts small late in 2026 and rolls out wider in 2027.

Why it matters to GCs

Cheaper AI to run means cheaper AI for you. When the cost to run these tools drops, the software you use for bids, schedules, and RFIs can add more AI features for less. Faster chips also mean less waiting on answers in the field.

TechCrunch·Aug 25·Read source →

LinkedIn