How to Build an AI Memory Layer for CRE
Every AI conversation starts from zero. Well, maybe not zero but close to it.
Everything we've published on AI in commercial real estate, newest first.

AI tenants account for 46% of active tech square footage, but nearly two-thirds of that demand is concentrated in San Francisco, Silicon Valley, and New York.
Every AI conversation starts from zero. Well, maybe not zero but close to it.
The unanimous code rewrite bars landlords from using competitors' non-public data to algorithmically set rents, fees, terms, or occupancy levels.
The paid membership gives agents visibility into client activity on Zillow and adds AI-driven prompts, listing signals, messaging, and branding tools.
The construction back-office startup closed a Series A led by Accel to add AI on top of the accounting, ERP, and project systems contractors already run.
A Dealpath survey finds 97% adoption but only half saving time, with fragmented data, not the models, cited as the main drag.
The luxury brokerage is rebuilding its tech around Google Cloud AI, spinning out a data company called Elius, and pledging three years of cost cuts.
A steady half-year headline hides a 57% quarterly drop, with a handful of mega-deals and debt financings propping up the totals.
The AI that pierces LLCs to find the real owner's cell phone. What it does, how it works, what it costs, and where the accuracy claims get loose.
David Moghavem built Trion Properties an MCP that reads years of its own deal data, and the plays are replicable, if you start logging now.
One Raven, backed by a $5M seed round led by Fifth Wall, keeps smart-home data on the home network, no cloud, no subscriptions, no behavioral data sold in the background.
Elise Beyond runs September 2 to 4 in Nashville, pitched as hands-on training led by the people who built the products and the operators running them.
A full-year operating budget that normally takes days can be assembled from a property's source documents in 1-2 hours using Claude, with every number source-cited.
A leveraged colocation operator is testing public-market appetite for AI infrastructure. How it prices is the first real read on how markets value leverage against the hype.
A three-product CRE platform that turns an OM into a proforma, searches 11 million parcels off-market, and runs lessor-side lease accounting. What it does, how it works, and what it costs.
The AI that builds your underwriting, picks your comps, ranks your buyers, and writes the deck. What it does, how it works, and where the human still sits.
QTS withdrew its appeal on the Prince William Digital Gateway, ending a three-year fight over 22 million square feet outside D.C. The project died on a notice technicality, not the merits, and that's the precedent.
401 megawatts of critical IT load, signed years before occupancy, functions like a 20-year ground lease with a hyperscale-grade credit tenant. Why CRE should read this as a real estate deal, not a tech headline.
Marc Rutzen turned a $25-a-month Google Sheet into HelloData, sold it to Grace Hill in under three years, and his playbook runs against almost everything founders assume.
The governor also wants to eliminate tax breaks for the facilities, marking a shift eight months after calling Texas the "epicenter of AI development."
The property-data giant takes a stake in an agent-owned Italian marketplace, with plans to route its commercial listings into LoopNet.
The homebuilding software company will expand beyond builders, naming the largest private building materials distributor as the first partner for its AI Estimating product.
An AI lending platform that shows where lenders are actually deploying capital, not where their marketing says they lend. What it does, who built it, and what it costs.
Andreessen Horowitz is leading the raise, which could precede an IPO as soon as 2027 and value the company at more than four times its 2022 buyout price.
A scheduled Claude agent that mines the rent comps you already have, scattered across old emails, dead OM attachments, and call notes, and logs them to a spreadsheet every Monday.
The world's first AI-driven CRE brokerage runs on a proprietary 2,470-deal dataset and a 24-foot hand-colored map. Here's what any broker can copy.
An AI underwriting platform built for the lender's side of the deal. What it does, where the claims get loose, and who it's actually for.
One of the country's largest apartment owners is putting a venture-backed generative-design tool into the earliest stage of its development pipeline.
A founder-led demo of the platform that makes 4,000 hours of daily local government meetings searchable, what it does, where it stops, and who it's for.
A Claude Code skill that convenes named retail CRE investors to deliberate, argue, run a pre-mortem, and reach a verdict on a real decision.
A founder-led demo of the all-in-one GP platform, what stood out, where it stops, and who it's actually for.
Why the real edge in AI-powered lending is not speed, it is the ability to defend every decision. And what that means for the sponsors and funds on the other side of the table.
Taylor Avakian turned his BOV, his CRM, and his content into an AI-powered edge, and the plays are replicable.
Hand it real skill and it produces a superhuman version of your work. Hand it confusion and it produces a faster, more confident, better-formatted version of being wrong.
Underwrite the same LA deal twice, cold, then with the firm's buy box loaded. The gap between the two outputs is the whole lesson in reliable AI underwriting.
Busy, impressive tools, the look of progress, and no deals. The AI mistake costing CRE teams, and the simple audit that catches it.
Most tools fail on the first question. The four-question filter I run before paying for any CRE AI tool.
Twenty years of deals are sitting unstructured in your email. That's the proprietary dataset no competitor can buy.