OpenAI turns Codex into white-collar software OpenAI shipped six role-specific Codex plugins, including plugins for investment banking and equity investing. The plugins combine tool integrations, instructions and context to enable Codex to perform specific tasks. Codex is shifting from an AI coding platform to an AI professional work, not unlike how Anthropic spun Claude CoWork out of Claude Code. A more accessible way to use ChatGPT’s coding platform for every day CRE tasks could be right around the corner. Plus, OpenAI launched the Codex Sites plugin, making it even easier to code, edit, and launch websites right from Codex. Be on the lookout for the Codex Sites tutorial in the Studio in the days ahead.
Google Earth turns Street View into a site-scouting tool Google Earth's AI data layers pull from billions of Street View images to map infrastructure like fire hydrants, storm drains, and utility access, then layer building footprints and land parcels on top, all available to Professional and Professional Advanced users. Google started rolling these AI layers out in 2025 but they’ve kept expanding the functionality through 2026.
Microsoft unveils 7 in-house MAI models at Build 2026 Microsoft rolled out seven of its own MAI (Microsoft AI) models, led by MAI-Thinking-1, its first reasoning model, and MAI-Code-1-Flash. Microsoft Copilot continues to be one of the most widely accessible AI tools to CRE professionals, and they’re now leaning into the development of their own AI models (instead of just relying on OpenAI.)
CRE tech firms continue to bet big on AI:
Cove expands its AI across whole CRE portfolios Cove, a building operations platform, just rolled out their AI layer to every client using the platform. They added new AI functionality in the form of three tools: Certificate of Insurance extraction that flags coverage gaps, work-order summaries, and AI-assisted tenant communications.
Entrata files to go public on the strength of its AI Entrata, the multifamily property management software firm, filed its S-1 on May 28 (NYSE: ENT), and the AI is the headline, not a footnote. Its agentic engine ELI automates leasing, payments, renewals, and maintenance, and 4 of the top 10 NMHC operators are already on the platform.
Legora buys Cadastral, a CRE-native AI deal platform An AI-legal platform, Legora, acquired Cadastral, a 10-person AI agent platform built specifically for CRE. Its agents are designed for the for commercial real estate workflows, from acquisitions and underwriting to due diligence, financial models, leases, refinancing, and zoning. Founded in 2024 and launched in 2025, Cadastral signed more than 50 firms in just over a year.
AI’s impact on office leasing:
AI growth delivers mixed forecast for office leasing Tech firms leased 11.5M SF in Q1, a decade-high 23% of all U.S. office leasing per CBRE, with AI companies clustering in a few buildings in San Francisco, Silicon Valley, and Manhattan. But Layoffs.fyi has already tracked 116,854 tech layoffs through early June, past the total for all of 2025, and Amazon alone is shedding about 14M SF. The AI boom seems poised to cut both ways in the office sector.
NotebookLM might be one of the most underutilized tools in CRE. It solves a problem that so many grapple as they’re using AI platforms: It’s built to not make stuff up. It’s AI that grounds itself only in the sources and information that you choose to provide it.
At least, that’s how it started. But NotebookLM has evolved since we ran our original tutorial last year. Your ability to synthesize the information in your Notebooks has gotten a massive upgrade, and they’ve built in features to help you actually source the intel you need for the task at hand.
Here’s the use-cases we walked through in last week’s refresh tutorial on NotebookLM:
1. Load the deal docs, interrogate them, extract the intel. Create your due diligence vault in a Notebook, and query the docs to pull the intel you need, or have Notebook compare them all against each other to suss out discrepancies. Better yet, you can now have Notebook extract data from your PDFs directly to a data table or spreadsheet.
What you get back depends on how you ask. The framework: AI Prompting for Commercial Real Estate: Tips That Actually Work.
2. Market analysis in minutes. Notebooks can now use Deep Research right inside your Notebook to find reputable information on whatever topic you need to digest. Use it to scour the internet for market reports on a market you’re expanding into, and create a detailed market analysis from the findings. Bonus: You can turn that market analysis into a podcast so you can listen to it on the way to your next meeting.
3. Map a client’s needs before you pitch. Same approach, different goal. Use Deep Research within Notebook to source publicly available intel on your potential client. Use the Mind Map feature to dig into the intel dump, and discover the things you should know before pitching that you didn’t even know you were looking for.
4. Build slide decks directly in your Notebook. You can generate slide decks using Google’s Nano Banana image generator directly inside your Notebook. Or, if a NotebookLM slide deck won’t cut it, have it draft the slide content for you based on your selected sources, and provide the content to the AI deck generator of your choice.
And that’s really just the start..
If your goal is to work with your sources and documents to find useful intel, and extrapolate it into formats that you or your team can easily digest, NotebookLM is your easy button.
Want the same grounded, sources-only recall inside Claude instead? That's a second brain: How to Build a Second Brain in Claude.
Find the entire NotebookLM tutorial update in the Studio.
Don’t miss our monthly live Claude Update Call in the Studio, where we go over Claude’s latest updates, and how to put them to work in our industry.
Claude Update Call
Date: Thursday, June 11th
Time: 12:15PM-12:45PM EST
Join before the call with your 7 day free trial.
Last Week’s Live Demo Session:
Last week, we had the team from Resimplifi join us for a live demo of their new Resimplifi Agent, an MCP that lets you query their nationwide commercial real estate dataset directly inside Claude, ChatGPT, or Perplexity. Resimplifi specializes in on-market listing data. Their database averages over 1,000 new CRE listings a day across the U.S., covering properties across every asset class: retail, multifamily, office, industrial, hospitality, mixed-use, and land. Sourced from national, regional, and local brokerages and verified weekly.
Now instead of scouring various websites for this data, it can all be available for you to query with the AI platforms you use every day.
Watch the entire live demo, see how you can leverage this nationwide CRE listing data, and get free access to the new MCP connector to test it out.
Find the full session in the Studio.