OpenAI Releases GPT-5.5, Bringing Company One Step Closer to an AI 'Super App' - OpenAI's newest model dropped April 23 with stronger coding, research, and document-handling abilities.
Should we in CRE care given how good Claude is? I think so.
We talked about that in the CRE AI Studio here.
Manager of 9,000 Apartments Testing New AI Tool to Find Tenants - Proptech Bonaventure ($2.8B AUM) is partnering with CorridorIQ, a UVA-founded startup using 15-20 data inputs like building permits, search trends, mortgage applications to predict migration patterns and target apartment marketing.
The trillion data points in the CRE industry are starting to become more accessible.
New Connectors in Claude for Everyday Life - Anthropic expanded Claude's connector system with 15+ new integrations turning Claude from a work chatbot into a daily life hub. Claude now suggests the right app based on conversational context, with no paid placements or sponsored answers.
Don’t know how to use connectors in Claude?
Watch our Claude Cowork video here.
Copilot's Agentic Capabilities in Word, Excel, and PowerPoint Are Now Generally Available - Microsoft flipped agentic Copilot to GA on April 22. It now takes multi-step actions directly inside spreadsheets, docs, and decks instead of just suggesting edits. Excel saw a 67% increase in usage per user. FINALLY!
I didn’t see any reason to advise people to use Co-Pilot in excel because it was more of a thinking partner than an actual agent like Claude in Excel was. Now that’s changed. So if you don’t have Claude, use this!
Google Launches Workspace Intelligence, Turning Its Productivity Suite Into an AI Office Intern - Google announced at Cloud Next on April 22, Workspace Intelligence is a unified AI layer across Gmail, Sheets, Docs, Drive, and Meet. Gmail now supports natural language search with AI Overviews, Sheets gets "prompt-based filling" that's 9x faster than manual entry, and Meet's Take Notes For Me now works for in-person meetings and calls on Zoom or Teams.
I still think Google is going to win the AI race (even if it’s through Anthropic, who they just poured another $40bn into…not bad).
Most CRE professionals know Claude as the chatbot you type questions into. But Anthropic has three distinct products — and each one unlocks a different level of AI capability for your real estate business. Here's the breakdown.
Claude Chat - Your AI Thinking Partner
This is the Claude most people know. The web and mobile chat interface at claude.ai. You type a question, upload a document, or paste some text and Claude responds. Simple, fast, no setup required.
When to use it in CRE: Claude Chat is your go-to for quick-hit tasks that live inside a single conversation. Summarize a lease abstract. Ask whether a cap rate makes sense for a Class B multifamily deal in a specific submarket. Draft an LOI response. Have it explain an SNDA clause in plain English. Brainstorm value-add strategies for a new acquisition. If the task starts and ends with text in a chat window, this is your tool.
Skill level: Beginner. If you can text, you can use Claude Chat.
Access: claude.ai (web), iOS, Android - free and paid plans available.
Claude Cowork - Your AI Assistant With Hands
Cowork is where Claude goes from conversationalist to operator. Available in the Claude Desktop app, Cowork mode gives Claude access to your actual files, your email, your calendar, Stripe, Slack — and lets it take action across all of them. It can read an Excel rent roll, build a PDF report, draft an email, and schedule a follow-up. It also has "skills", which pre-built playbooks for specific tasks like underwriting an OM or analyzing a T12.
When to use it in CRE: Cowork is for multi-step workflows that touch real files and real tools. Upload an offering memorandum and have it build an Excel pro forma with 5-year cash flows, debt analysis, and returns. Drop in a rent roll and get a professional PDF screening report with occupancy trends, loss-to-lease analysis, and lease expiration risk. Have it prep your weekly newsletter by pulling from your meetings, email, and community. This is the newsletter you're reading right now prepped entirely in Cowork.
Skill level: Intermediate. You don't need to code, but you'll get more out of it by learning to connect tools and use skills.
Access: Claude Desktop App (Mac and Windows).
New to Cowork? The setup that pays off later is here: How to Set Up Claude Cowork for Commercial Real Estate.
Claude Code - Your AI Developer
Claude Code is a command-line tool that lives in your terminal. You describe what you want built, and Claude writes the code, runs it, debugs it, and iterates autonomously. Think of it less as a chatbot and more as a junior developer who works at 10x speed and never sleeps.
When to use it in CRE: Code is for building custom tools that don't exist yet. One of our community members used Claude Code to build an entire smart call list generator with entity scaffolding, owner list filters, propensity scoring, and integration tests in about two hours. Other members have built property management dashboards, automated rent comp scrapers, tenant CRM apps, and custom underwriting calculators. If you've ever thought "I wish there was a tool that did X for my portfolio," Claude Code is how you build it.
Skill level: Advanced. You'll need comfort with the command line, but Claude Code handles the actual programming. Think of yourself as the architect and Claude is the builder.
Access: Terminal / CLI, VS Code integration.
The Bottom Line
Most CRE professionals should start with Chat, graduate to Cowork once they're ready for real workflow automation, and explore Code when they want to build something custom. You don't need all three on day one but knowing they exist means you can match the right tool to the right job instead of trying to force everything through a chat window.
We use all three pretty much daily, with Cowork being the tool we use the most.
Once you're working in Cowork, the next question is usually whether a task needs a skill or an automation. We broke that down here: Claude Skills for CRE vs. Automations.
Let me be clear: I am all-in on AI. I co-founded a company around it. I use it every single day. I believe it is the most important shift our industry has seen in decades.
And it's also creating a new generation of CRE professionals who have no idea what they're looking at.
I had a conversation recently with a real estate professor who's watching his students feed entire underwriting models into ChatGPT and submit whatever comes back. They're not checking the assumptions. They're not stress-testing the rent growth. They're not asking why the T12 shows a 40% spike in R&M in month eight. They don't know to ask because they've never had to build the model from scratch, line by line, at 2am before a Monday IC meeting.
That's the part that concerns me. Not that AI is wrong but that some of the young people using it don't have enough reps to know when it's wrong.
I spent 10+ years in multifamily asset management. I've walked units where the rent roll said "occupied" and the door was kicked in. I've caught tax reassessments that blew up an entire deal thesis that looked clean on paper. None of that shows up in a spreadsheet. None of that shows up in an AI output. That's judgment and judgment comes from reps.
So as more and more companies shift their focus from hiring to implementing AI, the lucky few who land roles have a dilemma.
They need to be really good at using AI because it’s expected. They also need to make sure that they don’t just take the experience they’re living (e.g. watching someone with experience model a deal from start to finish, build an investor deck and present it live) and try to replicate it in Claude.
They have to work hard to absorb the manual way to get really good at the automated way.
It’s like how we’re starting to teach building a skill in Claude, ChatGPT or other platforms. IF you start with asking the LLM to help you build the skill, the LLM doesn’t have the appropriate direct knowledge and will start guessing. But if you walk it through something first (e.g. analyze this T12 and follow these 10 steps…) and then ask it to build the skill, it already has the direct experience on the task and will build a much more accurate and valuable skill.
AI is a force multiplier. But you have to have something worth multiplying. If you've put in the years through walking the properties, building the models by hand, sat through the hard conversations with lenders and investors then AI makes you absolutely lethal. You move 10x faster with the same judgment you've always had.
But if you skipped the reps? AI just helps you be wrong faster and with more confidence.
So for those with the experience, use that as the leading force as opposed to relying on AI. And for those just starting out in CRE, get really good at using (and teaching) AI but absorb the lessons from that 10+ year vet the old-fashioned way and then feed it into your AI tool of choice.
You’ll be much happier (and more valuable) in the long run.