In competitive deals, the highest bid does not always win. Sometimes, the fastest one does.
If you have spent any time in acquisitions, you know how this works. Sellers do not optimize for price alone. They optimize for certainty of execution. A buyer who can move through diligence without drama, submit a clean offer quickly, and close on schedule will beat a higher bid that comes with financing contingencies, slow decision-making, or a reputation for retrading. The sophisticated buyers know this, which is why they structure letters of intent designed to lock out competitors early: short diligence windows, non-refundable deposits on day one, exclusivity clauses that take other bidders off the table before they have even finished reviewing the deal.
Speed is not a nice-to-have. It is the mechanism that converts deal flow into closed transactions.
Last month, a client called me after winning a multifamily acquisition deal they said they would have lost a year ago. The deal package was the usual mess: rent roll exported from a property management system with merged cells, a trailing twelve months of income-statement, contract PDFs that’s been triple scanned, a capex schedule living in an attachment called "final_v3_REVISED.xlsx." Close to a hundred pages across a dozen files that had to be made usable before anyone could underwrite it.
They submitted an LOI (Letter of Intent) inside forty-eight hours. The seller granted them exclusivity. No bidding war. No drawn-out process. They closed at a basis that made sense for their return targets while three other firms were still chasing down documents and reconciling spreadsheets.
How did they move that fast? They ran the deal package through an AI agent. The rent roll came out clean in minutes. The trailing twelve took longer because the chart of accounts did not map cleanly and a couple of line items were ambiguous, so the agent flagged them instead of guessing. But within about an hour they had a first-pass underwriting that actually tied out. Unit count matched. Gross potential rent matched what they expected within a small delta. The biggest anomalies had already surfaced. They could make a decision.
"We used to lose deals like this," he told me. "Not because we would bid lower. We would just be late."
How Speed Converts to Returns
Commercial real estate is a deal flow business. The constraint is rarely capital. There is plenty of dry powder sitting in funds waiting to be deployed. The constraint is attention.
A typical acquisitions team might see two hundred deals in a year. They seriously underwrite forty of those. They bid on fifteen. They close three. The economics of the business depend entirely on which three.
Now consider what happens at the bottleneck. Those forty deals that get seriously underwritten represent the team's capacity. An analyst can only build so many models, chase down so many documents, and reconcile so many discrepancies in a given week. Every hour spent on a deal that will eventually get killed is an hour not spent on the deal that will actually close.
Example One: CoStar Deal Pack to Underwritten Model
Most acquisitions teams start screening with CoStar. The problem is that information sits in a PDF, and the underwriting model sits in Excel. Someone has to bridge that gap manually before any real analysis begins.
The video above shows what happens when an agent handles that bridge. Upload a CoStar export, and the agent populates your underwriting template with unit mix, in-place rents, submarket comps, and preliminary assumptions for market rent, exit cap, and operating expenses. The output is not a finished model. It is a starting point that used to take an hour or two to assemble.
One client told me his team used to seriously evaluate about forty deals per year. After implementing this workflow, they evaluate over a hundred. Same team size. The constraint was never talent or judgment. It was the time required to get from "this looks interesting" to "here are the numbers."
From Passive to Proactive Deal Sourcing
Speed and capacity does not just limit how many deals you can close. It determines how you source deals in the first place.
Most firms operate in passive mode. They wait for brokers to send opportunities, react when something lands in the inbox, and hope the best deals come to them. Proactive sourcing, where you chase off-market deals, reach out to owners directly, and build relationships before properties hit the market, is how the best deals get done. Off-market transactions mean less competition, better pricing, and faster closes. But proactive sourcing requires capacity. You cannot pursue deals you do not have time to underwrite.
There is a second-order effect too. Brokers remember who responds quickly and who wastes their time. When you can turn around a serious analysis in forty-eight hours instead of a week, you become the first call on the next deal. When you consistently show up prepared with real numbers and specific questions, brokers start bringing you opportunities before they go wide. The firms that can move fast do not only win more competitive deals. They get access to better deal flow in the first place.
The firms generating alpha in this market are not the ones with the best models. Everyone has good models. They are the firms that can see more, move faster, and kill losers earlier. They bid on more deals with the same team. They get to conviction before their competitors. They deploy capital while others are still stuck in diligence.
This is where AI agents change the game. Not by replacing analysts, but by removing the bottleneck that keeps analysts from doing actual analysis.
Example Two: Every Deal Makes the Firm Smarter
Most firms have looked at hundreds of deals over the years. Every OM, rent roll, and underwriting model contains structured data about markets, pricing, and assumptions. But that knowledge lives in random Excel files scattered across shared drives, or nowhere at all. Each new deal starts from scratch. The firm isn’t getting wiser.
In TabAI, every deal you underwrite auto-ingests into a structured database. Over time, this becomes a living record of every deal your team has touched. The platform auto-generates pipeline views, comp benchmarks from your own history, and investor-facing materials. When you are raising capital, you can show LPs: "We have underwritten two hundred deals in the Southeast, fifty in South Florida. Here is how this acquisition compares to that universe."
This is where the alpha compounds. Every deal you look at, even the ones you pass on, makes the next decision faster and more defensible. Your conviction on bids comes from a live view of your own comps, not memory. Your LP conversations show a systematized process, not ad hoc Excel screenshots.
Most tools help you fill in the model. The real edge is owning the knowledge graph that builds itself and keeps getting smarter.
Agents vs ChatBots
Most applications of artificial intelligence in real estate start with language tasks. Summarize this offering memorandum. Draft a paragraph for an investor letter. Answer a question about capitalization rate trends. These are useful, but they do not solve the core problem. They are chatbots, not agents.
An agent is different. Rather than responding to a single prompt, an agent executes a multi-step process with checkpoints. In the context of underwriting, this means repeatedly doing four things: identifying what needs to happen next to reach a lender-ready output, extracting key information from source documents, verifying those facts against each other, and producing the specific deliverables that investment committees actually review.
The distinction matters because underwriting is not a single task. It is a workflow. Documents arrive in inconsistent formats. Data needs to be normalized before it can be analyzed. Assumptions need to be checked against multiple sources. The final model needs to tie to the final memo. An agent handles this entire sequence, surfacing issues for human review rather than waiting for a human to discover them.
On that Multifamily deal, the agent parsed the rent roll into a clean table in a few minutes. It normalized the trailing twelve-month operating statement into a standard chart of accounts. It flagged a couple of line items that needed clarification from the seller, but nothing that changed the thesis. It populated the underwriting template with in-place assumptions and generated a summary the team could review immediately.
The process that typically takes an analyst three to five hours took under an hour. But here is what mattered more than the time: they had conviction. They knew their numbers were solid. They moved fast and won the deal.
The Industry is Getting Faster
JLL, CBRE, and MSCI are all building AI platforms targeting underwriting workflows. CoStar has been integrating AI across their products for years. These are not experiments. They are production investments from firms with direct visibility into where their clients spend time.
The implication is that speed and rigor are becoming table stakes. The firms that adopt early will see more deals with the same headcount. The firms that wait will find themselves competing against teams that move thrice as fast.
But agents do not replace what actually matters. They cannot tell you that a broker has a reputation for hiding problems, or that a particular lender is hungry for this asset type right now. They cannot tap a network for off-market deals or read the room in a negotiation. The best investors are the ones who know people and have built trust over decades. Agents free up more time for that work.
The Need for Speed
The window for competitive advantage will not stay open forever. As more firms adopt these tools, the edge shifts from "faster than competitors" to "keeping pace with the market." The firms that move early will have compounding benefits: better processes, deeper institutional knowledge about what works.
Commercial real estate will not become fully automated. But the work that supports investment decisions is already changing. The firms using these tools are seeing more deals, moving faster, and deploying capital while their competitors are still reconciling spreadsheets.
Andrew Siah is Co-Founder and CEO of New York AI Labs, which develops TabAI, an AI-powered workspace for complex finance. He holds a degree in physics from Columbia University and previously worked in quantitative trading. He has published AI Research Papers at NeurIPS and ICLR, and peer reviewed works at ICML. He was a guest speaker in Professor Tianyi Peng's Generative AI class and will be presenting in Earle W. Kazis and Benjamin Schore Professor of Real Estate Stijn Van Nieuwerburgh's Real Estate Analytics class in Spring 2026. Andrew can be reached at [email protected].