Inventory Valuation Methods

Still choosing properties the old way? The market moved on yesterday. From Asia to the Americas, real estate is being redefined by algorithms, not anecdotes. Investment decision-making is no longer just about price trends and location. Factors like energy infrastructure, tenant demand, and building performance are being decoded in real time to hep RE investors—using AI, LiDAR, IoT, and predictive analytics. In one standout example, a city initiative in Calgary, Canada, used 3D building models and advanced data tools to help residents estimate solar potential on rooftops. The result? A dramatic rise in solar installations and a blueprint for how data can accelerate infrastructure adoption. But it’s not just residents driving this shift. Developers and investors are already using the same technologies to guide large-scale decisions—whether it’s optimising energy consumption, increasing occupancy, or identifying high-performing assets long before the market catches on. The new paradigm is here. Real estate is fast becoming a data-first industry. And now, generative AI (Gen AI) is sharpening the edge—from analysing lease documents at scale to visualising human-centric interiors optimised for light, movement, and acoustics. Imagine asking: - “Which 25 warehouse assets will outperform over the next decade?” - “Design tenant spaces based on actual behaviour patterns—and optimise for comfort, daylight, and energy use.” Gen AI doesn’t replace your investment instincts. It enhances them—by delivering faster insights, personalising tenant experience, unlocking new revenue streams, and shortening decision cycles. At CBRE, we’re equipping clients with cutting-edge data analytics platforms and AI tools that turn real-time information into real-world value. From portfolio benchmarking to dynamic planning and predictive modelling, our technologies are designed to help you lead, not follow. The tools are here. The use cases are proven. The competitive advantage? Still up for grabs. Are you using analytics to simply observe the market—or to outpace it? #RealEstate #PropTech #DataAnalytics #AI #GenAI #SmartInvestment #CBRE #Innovation #DigitalTransformation

AI real estate valuation - an updated guide to VIBE valuation.   I've posted about using Claude and DeepResearch for valuation before (link in comments), but we now have three new models that can do this: ChatGPT o4 mini & o3 as well as Gemini 2.5 – drop the PDF brochure and ask.   I’m using a brochure for 89-91 Gresham Street (from Sep 2022).   Last month #ClaudeAI and ChatGPT #DeepResearch valued the property at around £13-14m with a unlevered IRR of around 9%. Today, the smartest models we have all suggest a price of around £11m reflecting a more pessimistic outlook.   ChatGPT o3: gave me a detailed analysis with simple prompting. It was ruthless to call the property “core+” and ask for a 9.6% unlevered (12% levered) IRR. It was generous with leasing and growth assumptions so the price was £10.42m. ChatGPT o4 mini high: took a very numbers-focused approach, assumed that a core property in London would need a 7% unlevered IRR (calculated 9% with debt using the WACC formula!) and with simple assumptions (and some follow up promoting) valued the cash flows at £11m.   Gemini: was very similar to o4 and took a very quantitative approach focusing on the numbers rather than understanding the property. It was optimistic with growth and leasing assumptions, but assumed higher OpEx. I had to prompt to get a good answer. Eventually it asked for 7% unlevered IRR (10% with debt) and valued the property at £11.32m.   Some reflections:   1. It’s remarkable how close the three models were. While the value is lower than last month, it’s still consistent across AI models that write valuation models in different ways.   2. AI models make assumptions and cash flow modelling decisions – they are drawn from a distribution, so running the same valuation prompts in the same model does not always give the same numerical outcome. However, if you run the same prompts 30 times and take the average, you can replicate the average by running them again. If you want reproducibility, ask AI for code.   3. Every time the "valuation date" was the date at which the brochure was produced and not today. This highlights the fact that they focus on getting the numbers right rather than thinking about the nature of what they do. ChatGPT o3 is better at this than others but not perfect.   We still need humans to drive and supervise the process, but Ai can speed things up.   Controversial opinion: I think getting AI to do first pass-valuations is a good idea. If done well, it can be helpful even without detailed checks. I wouldn’t trust a single run of a single model, but 10 runs on 4 different models would be an interesting signal. ✨ Following the AI naming convention it's "VIBE valuation".   I think it’s important, so I’m adding this to our Exec online course in June (link in comments). For chat histories with results, link is in comments (you have to sign up to the newsletter). #AIValuation #PropTech #RealEstateAI #VIBEValuation #CRE #CommercialRealEstate #AIinFinance

Real estate isn’t buy low, sell high. It’s buy right, based on features. 🏡📊 I built this Real Estate Sales Power BI Dashboard to help investors/analysts quickly spot what actually drives price + profitability from raw property transactions without drowning in rows of data. What this dashboard answers: • Which bedroom counts give the best value (diminishing returns after ~7–8) • How size impacts price and how waterfront breaks the normal rules • Whether waterfront homes are worth it (only ~0.75% listings, but huge premium) • What drives valuation the most: grade + view + bedrooms (Decomposition Tree) • Month-wise price movement (seasonality + volatility) • Where most deals happen: $200K–$600K dominates volume, while $1M+ drives value What I practiced in my analytics journey: • DAX measures for Sales, Profit, ROI, Price/Sqft • Decomposition Tree to explain why price changes • Building a clean, decision-focused dashboard for business use 🔗 Live Dashboard: https://lnkd.in/dsiuUSVD #PowerBI #DataAnalytics #DataVisualization #RealEstateAnalytics #BusinessIntelligence #Dashboard #DAX #PowerQuery #DataStorytelling #AnalyticsJourney #LearningInPublic

What makes a property truly comparable? Six years ago, we built a model to find out. No #ChatGPT. No #GenAI hype. Just frustration with how comps were selected — often because they matched the outcome, not the asset. Same valuation. Same rent. Same expense levels. Selection bias dressed up as methodology. And if you did try to do it right? It took forever. So we built a machine learning model to score similarity — and JUST got the patent: US 12,254,030 B1. Key takeaways: • Most of the predictive power came from traditional features — renovation spend, unit mix, overall property quality — though distance to Starbucks was the one untraditional variable that consistently mattered • NYC is built different (no surprise) — it needed its own weighting model for almost everything • Might’ve been early — and better suited for appraisers, owners, or origination teams than agency underwriters Plenty of platforms do this today — HelloData’s work stands out. Tools have improved, but the challenge of accurately, objectively, and efficiently identifying comps is still very real. Credit where it’s due — the co-inventors, the real brains behind this: Steve Guggenmos Eugenia Yerukhimovich Jun L. Dina Guo, Ph.D. Biao Yang, Ph.D. @Patrick Chu And early believers who helped get it off the ground: Freddie Mac Multifamily Christine Halberstadt Raj Bector Filicia Davenport Curious — what’s your go-to comp tool today? #comps #proptech #machinelearning #XGBoost #CREtech #AIpatents #AIIP #valuation #CRE

⚡️ I Used AI to Analyze a 60-Property Portfolio in 30 Minutes — Here's How It Changed Everything ⚡️ In real estate investing, the real value comes from asking the right questions and structuring smart deals — NOT spending days crunching numbers or gathering property data. Here's how I completely transformed my workflow: 1️⃣ Instant Portfolio Data — I pulled full details on 60+ properties in seconds using HouseCanary. No more endless MLS searches, browser tabs, or Excel imports. 2️⃣ Fast, Smart Analysis — I used modern analytics to instantly size up each property's value, identify strengths and risks in the portfolio, and determine what I’d pay. 3️⃣ AI-Powered Investment Memo — With all the data in place, I ran it through AI to generate a comprehensive investment memo in seconds — outlining ROI, risk, and deal viability. ✅ What used to take days of grunt work — comping, formatting, spreadsheet modeling — now takes less than half an hour. 💡 AI won’t replace the investor. But it eliminates the repetitive parts, so I can spend time doing what actually matters: - Spotting hidden risks - Thinking through the best structure - Crafting a winning deal The result? Better, faster investment decisions — and a sharper edge in a competitive market.

I asked three AI models to value the same Nicosia apartment. Nine times. Here's what happened. Same prompt. Same property: a 75 sqm, 2-bedroom apartment in Acropolis, built 1995, average condition, 3rd floor, near the Central Bank of Cyprus. Three models. Three runs each. Nine answers. The spread: €145,000 to €220,000. That's a €75,000 gap — roughly 50% of the lower estimate — for the same property on the same day with the same brief. A few observations worth sitting with: 1. None of the models had transaction data. They averaged over public listings. In Greece, our own data shows residential listings are inflated by ~24% relative to actual transactions. In Cyprus, public listing data is similarly noisy. AI without structured, verified inputs is a confident-sounding averaging machine over whatever happens to be on the open web. 2. None of them inspected the property. No lift check. No common-area assessment. No title verification. No EPC. No parking confirmation. Each of these is a 5–15% swing factor. 3. None of them could be signed. A RICS Red Book valuation carries professional indemnity, regulatory standing, and admissibility to a credit committee or a court. An AI output carries none of that. When something goes wrong on a €20m portfolio, someone has to be on the hook. 4. Most clients can't legally use them anyway. Pasting a borrower's collateral details into a US-hosted foundation model is not a preference question for a regulated bank, insurer, or fund — it's a GDPR, DORA, and EBA outsourcing problem. The point isn't that AI is bad. It's genuinely useful, and at Ask Wire we use it daily. The point is that AI is only as good as the data layer underneath it and the framework around it. The firms winning the next five years in Cyprus and Greek real estate won't be the ones who replaced their analytics provider with ChatGPT/ Gemini/ Claude. They'll be the ones who plugged a regulated, locally-grounded data infrastructure into their AI stack — with human accountability where it legally and commercially matters. AI changes how analysis gets produced. It doesn't change where the data comes from, who signs off on it, or whether you can defend the number in front of a regulator. That's the layer we're building.

You can’t outsource understanding. A recent "What Would BK Do?" question got right to the heart of something I’ve spent decades working on. Daniel asked how we normalize land comps over time, especially in a city like New York where zoning changes, FAR shifts, tax abatements, and liquidity cycles are constant variables. In other words: How do you make sense of land values across multiple market cycles without fooling yourself? The short answer is discipline. In the 41-year Manhattan land study, we don’t try to outsmart the data with assumptions. We start with straight averages, but we disaggregate everything into buckets that actually behave differently: rental residential, condo residential, office, hotel, and a miscellaneous category. Lumping those together would be like averaging a peach, a bowling ball, and a two-by-four. The number would be meaningless. All analysis is done on an as-of-right basis. Zoning changes may increase density, but they don’t automatically increase value on a price-per-foot basis. Value still comes from rents, end-user pricing, capital markets, and broader economic forces. From there, we look at how land values fluctuate against a basket of macroeconomic metrics: interest rates, inflation, equity markets, commodities, lending conditions. Not to predict, but to understand what tends to matter when direction changes. That’s the answer. But, the more important lesson is this: You can't outsource understanding. If you rely solely on third-party data, you’re inheriting someone else’s assumptions, errors, and shortcuts. Over a long career, that’s dangerous. Building your own datasets, verifying them transaction by transaction, and using the same methodology over decades creates something far more valuable than a clever forecast: perspective. Markets change. Cycles repeat. The brokers and investors who last are the ones who know why the numbers moved, not just that they moved. That’s what having your own data really gives you. #WhatWouldBKDo #NYCRealEstate #BKREA

High income in real estate should not feel like guessing with 7-figure chips. But many smart investors are still playing analog games in a digital world. I have seen tech leaders use AI at work to drive huge outcomes, then buy property with gut feel, old comps, and a broker pitch. The issue is not effort. It is that the strategy is stuck in a pre-AI playbook. ⚠️ The real threat is not that AI will replace investors. It is that AI will quietly reward the investors who use it while everyone else keeps donating returns. ❌ Many portfolios run on scattered spreadsheets, late data, and emotion, so risk hides in plain sight and opportunities slip by before anyone notices. How to use AI in real estate to build durable Legacy Wealth 1️⃣Move from gut feel to real data. ↳ Let AI read jobs, rents, migration, and sentiment so each decision is grounded in probabilities, not vibes. 2️⃣Use AI to find deals before the crowd. ↳ Set rules for growth and let AI scan thousands of markets to surface emerging neighborhoods 6–18 months early. 3️⃣Tighten pricing with smarter valuations. ↳ Blend comps with micro data and macro trends so offers land in a tight band where upside is real and overpaying is rare. 4️⃣Stress test every deal before money moves. ↳ Run rate, rent, vacancy, and exit scenarios so weak deals die on a screen instead of in your portfolio. 5️⃣Turn portfolios into live systems, not static lists. ↳ Connect rent, expenses, and tenant data so AI can flag underperforming assets and hidden NOI wins. 6️⃣Automate operations without losing control. ↳ Use AI leasing, follow ups, and predictive maintenance to cut vacancy and repair costs while keeping the human touch. 7️⃣Make financing speed a strategic weapon. ↳ Feed clean data into AI-enhanced underwriting so lenders move faster and your offers win on certainty, not just price. 8️⃣Build ethics and governance into every model. ↳ Audit data and outputs for bias, document how AI is used, and keep humans responsible for decisions that affect people’s lives. 9️⃣Plug AI into a long-term Legacy Wealth plan. ↳ Align AI insights with entity design, tax strategy, capital allocation, and a clear path to a $5M+ portfolio that outlives any single cycle. When AI and real estate work together, investing stops feeling like jumping from deal to deal and starts looking like an operating system for Legacy Wealth. The question is no longer “Will AI change the game?”. It is whether it will quietly grow your balance sheet or someone else’s. ❇️ Your next step 👉 If you’d like the full framework + examples read the blog in the comments. 👉 Want to stress test your own portfolio. Book a free 1:1 call to design a personal wealth strategy and map how to turn $100K-$1M+ of tax drag into $5M+ Legacy Wealth over the next decade:  https://lnkd.in/gwa5gqZG . Enjoy this? ♻️ Repost, follow Ravi Katta, and check the link in bio for more content and resources on building Legacy Wealth.