Why I'm a Commercial Real Estate Broker Who Codes

Licensed CRE broker. Also writes Python, ships iOS apps, runs 54 algorithmic trading bots. Most people in commercial real estate couldn't explain what an API does if you spotted them the acronym — and that's not a knock on anyone. It's an observation about the single fattest competitive advantage I've stumbled into across every industry I operate in.

McKinsey ranks construction and real estate among the least digitized industries on the planet — under 15% as digitized as tech and media (McKinsey Global Institute, "Digital America," 2015 — still cited in their 2024 updates because, remarkably, the gap hasn't closed). From a different vantage point, JLL's 2025 survey paints a similarly lopsided picture: 92% of CRE occupiers have started or plan AI pilots. Only 5% claim they've nailed all their AI goals. A mere 33% of the workforce feels adequately trained.

Massive market. Primitive tech. That's where I operate.

The CRE Industry's Technology Problem

Walk into most commercial brokerages and you'll encounter the same stack you'd have found in 2015. Excel spreadsheets ricocheting around via email. Deal tracking living inside someone's skull. Property data yanked manually from CoStar or LoopNet, one search at a time. Financial models rebuilt from zero every single transaction.

Why hasn't this changed? Part of it is structural inertia, part of it is incentive misalignment — but the numbers tell the real story. Sixty percent of CRE firms still run on legacy technology infrastructure (Deloitte, 2024 "RE-generative AI" report). AI-skilled job postings in real estate grew 64% in 2022, then another 58% through August 2023 — but those roles land at PropTech startups and institutional REITs (real estate investment trusts — publicly traded companies that own and operate income-producing properties), not at local brokerages where deals actually close. The typical CRE broker's tech stack? A phone, a car, and a listing database login. Seriously.

Now, the big three are spending. CBRE's AI-based Nexus platform manages over 1 billion square feet across 20,000+ client sites. JLL launched JLL GPT in 2023. Cushman & Wakefield rolled out AI+ and PRISM, powered by GIS (geographic information systems — mapping and spatial analysis technology) for brokers. Enterprise-grade platforms at enterprise-grade pricing. None of that trickles down to a solo broker running their own book.

So I build my own.

What I Actually Build for My Practice

Every tool I've written starts the same way — I kept slamming into a friction point until the annoyance outweighed the effort of coding a solution.

Deal underwriting automation. A standard CRE acquisition chews through dozens of analyst hours on data extraction alone — companies like Buildout and Blooma have built entire businesses around this bottleneck. I wrote a Python script that pulls rent rolls, calculates NOI (net operating income — total revenue minus operating expenses, excluding debt service), applies cap rates from comparable sales, and spits out a sensitivity analysis across five scenarios. Under a minute. That same analysis done by hand? Four to six hours hunched over a spreadsheet, hoping you didn't fat-finger a cell reference.

Market data pipelines. The old way: log into CoStar, run a search, export a CSV, reformat the whole thing for a client deck, realize you forgot a filter, start over. I automated the entire flow — market comps, vacancy rates, recent sales, demographic data. The pipeline runs on the same launchctl-based scheduling system from my automation stack. Different domain, identical plumbing.

Client reporting. Monthly market updates that used to devour two hours of my morning now take ten minutes. Data aggregation, formatting, chart generation — all handled before I've finished my coffee. My job becomes the insight layer. Not wrangling CSVs.

Here's the thing nobody in PropTech wants to say out loud: software companies automated this stuff for other industries decades ago. CRE hasn't caught up because the brokers who understand the business don't code, and the developers who could build the tools have never brokered a deal. I happen to sit in both chairs — which felt accidental at first but turns out to be the whole point.

The Deal Analysis Edge

Where does coding pay for itself fastest? Deal screening. Not even close.

The financial modeling fundamentals I use for M&A translate cleanly into CRE deal evaluation — but I can run them at scale, across dozens of opportunities simultaneously, while a traditional analyst is still formatting their first spreadsheet.

Morgan Stanley (2025) puts it at 37% of real estate tasks automatable, worth $34 billion in operating efficiencies by 2030. Brokers and services get the biggest lift: a projected 34% increase in operating cash flow through automation — the highest of any CRE subsector.

Here's a simplified version of how I screen a deal before committing to full underwriting:

# CRE deal quick-screen — runs before full underwriting
def quick_screen(noi, asking_price, market_cap_rate, loan_rate, ltv=0.70):
    implied_cap = noi / asking_price
    debt_service = (asking_price * ltv * loan_rate) / 12
    monthly_noi = noi / 12
    dscr = monthly_noi / debt_service if debt_service > 0 else float('inf')
    cash_on_cash = (noi - (debt_service * 12)) / (asking_price * (1 - ltv))

    return {
        "implied_cap_rate": f"{implied_cap:.2%}",
        "spread_vs_market": f"{(implied_cap - market_cap_rate)*100:.0f} bps",
        "dscr": f"{dscr:.2f}x",
        "cash_on_cash": f"{cash_on_cash:.2%}",
        "verdict": "ANALYZE" if implied_cap > market_cap_rate and dscr > 1.25 else "PASS"
    }

What's happening under the hood: the screen calculates DSCR (debt service coverage ratio — how many times the property's income covers its debt payments, where 1.25x is the standard lender minimum) and cash-on-cash return at a default 70% LTV (loan-to-value — the proportion of the purchase price financed by debt).

Thirty seconds. Kills most deals dead before I invest a single hour of underwriting time. A traditional broker reads the full offering memorandum (OM — the marketing package a seller's broker distributes to prospective buyers), grinds through the numbers by hand, and discovers hours later that the cap rate (capitalization rate — NOI divided by purchase price, the CRE equivalent of a P/E ratio) doesn't pencil. I know before my coffee gets cold. Same math, same financial principles — the difference is raw throughput. More deals evaluated, faster, with better opportunities surfaced while slower competitors are still muddling through their first read-through.

Why Most Brokers Won't Code (And Why That's My Advantage)

The barrier isn't ability. It's incentive structure.

Brokerage compensation runs entirely on commissions. Every hour spent learning Python is an hour not prospecting, not touring properties with clients, not grinding toward a close. The return on coding stays completely invisible until the system is built and humming — all upfront investment, delayed payoff, with no guarantee the broker sticks with it long enough to see the compounding kick in.

And that's precisely what makes it a durable moat. Anything hard to replicate and slow to pay off keeps competitors out indefinitely. Relationships remain the traditional CRE moat — and they still matter enormously, maybe more than ever in a market drowning in digital noise. But relationships combined with technical capability? I can count on one hand the brokers in my market who bring both to the table.

JLL's 2025 survey reinforces the point: 92% of CRE companies pilot AI, but only 33% of their workforce feels trained to use it. The tools are flooding in. The skills aren't following. My edge isn't the tools themselves — it's being the person who builds, customizes, and operates them without waiting for a vendor to ship a feature request.

PropTech Is a $47 Billion Market — Most Brokers Can't Access It

Precedence Research puts the global PropTech market at $47.08 billion in 2025, heading for $185.31 billion by 2034 at a 16.40% CAGR (compound annual growth rate). VC money poured in — $16.7 billion in 2025, up 67.9% year-over-year per CRETI (Center for Real Estate Technology & Innovation).

Source: CRETI 2024 and 2025 PropTech Venture Capital Reports

Most brokers experience PropTech as something that happens to them — another platform to learn, another subscription to pay, another sales rep promising "seamless integration" that turns out to be anything but. I treat it differently. Tool doesn't exist for a workflow I need? I write it. Existing tool won't plug into my stack? I build the bridge myself, usually in an afternoon.

McKinsey estimates gen AI alone could unlock $110-$180 billion in value for real estate (McKinsey Global Institute, June 2023). But here's what the consulting decks leave out: that value doesn't materialize because you bought a SaaS subscription and attended the onboarding webinar. It shows up when someone who understands both the deals and the code applies AI to real workflows — underwriting, market analysis, lease abstraction, tenant screening, portfolio optimization. The gap between "AI-enabled" on a pitch deck and actually using AI to close deals faster is enormous, and almost nobody in CRE is bridging it.

I'm not waiting for PropTech to catch up.

What Coding Actually Teaches You About Deal-Making

Debugging is due diligence. When I debug a trading bot at 2 AM — tracing an error from symptom to root cause, checking assumptions, isolating variables, testing hypotheses — it follows the exact same logic as good due diligence. Don't take the OM at face value. Verify the rent roll line by line. Trace NOI from source documents. I've caught discrepancies in deal packages that experienced brokers missed entirely, and I'm convinced it's because I read financial documents the way I read code. Assume nothing. Verify everything.

Automation rewires how you think. Most brokers operate reactively — deal lands on their desk, they analyze it, they respond. Building automation forced me into systems thinking: what's repeatable, where do bottlenecks hide, what genuinely needs my judgment versus what can be templated and forgotten? That mindset bleeds into pipeline management, client relationships, deal flow. Everything, really.

Version control is documentation discipline. Every line of code I write gets tracked, versioned, made recoverable. Same rigor applies to my deal files — every version of every analysis, every communication thread, every amendment. Client asks about a revision from three weeks ago? Found in seconds. Meanwhile, most brokers are excavating email threads and praying they didn't delete the attachment.

I build in public because writing about these cross-domain connections forces me to sharpen them. The documentation habit that produces posts about my trading bot operations is the same habit that makes me document deal processes — and both improve from the discipline. Strange how that works.

Building the Bridge: Broker Credibility Meets Technical Capability

CRE doesn't need more developers parachuting in from the outside to build PropTech. It needs practitioners building technology from the inside — people who've actually sat across from a tenant, negotiated a lease amendment at 11 PM the night before closing, walked a vacant warehouse trying to visualize three different layouts for three different prospective users.

US CRE investment hit $171.6 billion in Q4 2025 alone — up 29% (CBRE Quarterly Figures). JLL's 2023 survey says 80%+ of CRE investors and developers plan to increase tech spending. ICSC launched ICSC+PROPTECH in March 2026, wiring CRE decision-makers directly to technology founders. The money is chasing this intersection hard.

But most people building CRE technology have never brokered a deal. Never walked a property with a client who's trying to visualize their business in the space. Never fielded a panicked call from a tenant whose HVAC failed on a Friday at 5 PM. They build tools that demo beautifully at conferences and fail spectacularly in practice because they've never done the actual work.

I build from the opposite direction. Every tool starts with a real problem from an actual transaction — something that burned me once, or twice, or enough times that the frustration finally boiled over. Code serves the brokerage. Not the other way around. And the brokerage informs the code in ways no product manager interviewing brokers from outside could ever replicate.

What does this look like in practice? A single system where both sides compound. Financial modeling feeds the deal analysis scripts. Scripts surface better opportunities. Better opportunities build the brokerage. The brokerage generates problems worth solving with code. Flywheel.

If you're a developer curious about CRE, or a broker wondering whether coding is worth your evenings and weekends — the technology gap isn't closing. Not soon, anyway. The people who bridge it will hold an advantage that pure relationship brokers and pure technologists simply cannot touch.

Frequently Asked Questions

Do you need to know how to code to be a successful commercial real estate broker?

No — most successful CRE brokers have never written a line of code. Relationships, market knowledge, and deal execution remain the foundation. But coding creates tangible advantages: Morgan Stanley (2025) projects a 34% increase in operating cash flow for CRE brokers and services through automation. The edge isn't replacing traditional skills — it's amplifying them. Analyze faster, serve clients better, screen more deals in the same number of hours.

How is technology changing commercial real estate in 2026?

Fast. JLL's 2025 survey: 92% of CRE occupiers have started AI pilots, but only 5% have hit all their program goals. PropTech VC hit $16.7 billion in 2025 (CRETI), up 67.9% year-over-year. The tech is arriving — the training isn't. Only 33% of CRE professionals feel adequately AI-trained (JLL, 2025).

What programming language is most useful for commercial real estate?

Python, and it's not close. It handles financial modeling, data analysis, API integrations, and automation — the four highest-value CRE applications. I use the same Python infrastructure for CRE deal analysis that powers my 54-bot automation stack. Libraries like pandas, numpy, and matplotlib map directly to the data-heavy analytical work brokers do daily.

How big is the PropTech market?

The global PropTech market reached $47.08 billion in 2025 and is projected to grow to $185.31 billion by 2034 at a 16.40% CAGR, according to Precedence Research. PropTech VC investment hit $16.7 billion in 2025 (CRETI), with $3.2 billion going specifically to AI and automation in 2024. McKinsey estimates generative AI alone could generate $110-$180 billion in value for the real estate industry.

Can a solo broker compete with CBRE and JLL on technology?

Not on enterprise platforms — CBRE's Nexus manages over 1 billion square feet across 20,000+ sites. But a broker who codes can build custom tools that solve specific workflow problems the enterprise platforms don't address. The advantage isn't scale — it's specificity. My deal screening script filters opportunities in 30 seconds because I built it for exactly how I evaluate deals. No enterprise platform offers that level of customization at the individual broker level.

Is learning to code worth the time investment for a real estate professional?

The opportunity cost is real — every hour coding is an hour not prospecting. But the leverage compounds faster than you'd expect. Deloitte found CRE AI job postings grew 64% in 2022 and 58% more through 2023 — the industry is moving toward technical fluency whether individual brokers adopt it or not. Start with whatever you do most often: market data pulls, client report formatting, deal screening. Ten hours of coding saves hundreds over a career.

CRE and code is where I find the best problems — and the widest moat. The 54 trading bots across five markets, the automation infrastructure powering everything, the financial modeling bridging M&A and deal evaluation, the context engineering powering my AI workflow, the vibe coding framework for knowing when to trust AI output, the agentic engineering methodology for working with AI in each session, agentic AI across all four ventures, what breaks when you automate everything, and building publicly — it's all connected. Explore what I'm working on or get in touch.