Much of the AI conversation in the commercial real estate world centers on brokers, landlords, and the buildings themselves. But there’s a parallel transformation happening one step removed, inside the banks, lenders, and finance teams that fund the deals we work on. It’s a shift that matters to those on the ground because it’s changing the speed of loan approvals, how properties are valued, and the level of scrutiny a deal gets before it closes.
Why banks are leaning into AI now
Part of this urgency stems from pressure. An estimated $2.2 trillion in CRE loans will come due for repayment or refinancing over the next few years — that’s a real concern for financial institutions, especially if defaults or delays begin to accumulate. Combine that influx with tightening margin pressure and elevated funding costs, and banks have a strong incentive to sharpen their risk approach, accelerate decision-making, and adopt a leaner back-office model.
The financial upside of this transition? Huge. According to one industry leader, financial institutions that fully embrace AI could see up to a 15-percentage-point improvement in their efficiency ratio, split between revenue growth and cost reduction.
One institution cited in this article saw a 40% drop in the cost of verifying commercial banking clients from AI-driven onboarding tools alone. For commercial real estate professionals, that type of back office efficiency gain trickles down to a smoother, faster onboarding experience and a compressed timeline to sit at the closing table.
Banks continue rethinking their own office footprints as they embrace AI and onboard more tech talent. The financial services industry saw a 99% increase in computer and information systems manager roles between 2019 and 2025. Meanwhile, traditional operational roles like tellers and loan officers declined.
Because banks are competing directly with Silicon Valley for top-tier data scientists, financial firms are increasingly treating premium office location and collaborative workspace design as a talent-recruitment tool. This talent war is reshaping demand and driving flight-to-quality leasing trends in major global banking hubs, including New York, London, Dallas, Atlanta, Toronto, and Warsaw.
AI and property valuation: A real shift from the old AVM
Automated valuation models (AVMs) have existed in banking for years, but they’ve always had a reputation as a blunt instrument. They were useful for a quick ballpark estimate, but lacked the nuance to carry real underwriting weight for complex commercial assets.
What’s evolving now, however, is the sophistication and velocity of the inputs. AI-driven underwriting platforms can pull in economic indicators, demographic shifts, tenant behavior data, and live market comparables simultaneously. These platforms generate predictive risk models that update in something close to real time. Tools in this space can flag undervalued assets, forecast rent and absorption trends, and incorporate foot traffic and tenant movement data that a traditional appraisal can’t capture.
Additional research points to a similar shift on the lending side: AI can continuously monitor volatility in a given CRE market and model how that volatility affects property valuations instantly. This data feeds directly into a bank’s risk assessment rather than waiting for a periodic reappraisal.
What’s happening in underwriting and credit decisions
MIT Sloan’s research shows how AI is impacting the lending side. Before a human underwriter reviews the file, AI systems can now:
- Extract data directly from loan application documents.
- Verify income and financial records.
- Flag inconsistencies.
- Generate a preliminary credit risk assessment based on historical portfolio performance.
There’s still a human in the loop who signs off on the final decision, but much of the document chasing and first-pass analyses that took days can now happen in minutes. This automation frees human underwriters to focus on the art of the deal, evaluating sponsor character, structuring complex terms, and managing client relationships.
Those working in commercial lending specifically can use generative AI to parse years of financial statements to spot irregularities, evaluate market trends tied to a borrower’s business, and estimate collateral value. Lenders can run “what-if” scenarios to see how a new loan would affect a bank’s overall risk exposure if the borrower defaulted.
Treasury, cash flow, and fraud detection
Less glamorous than loan approval — but important for managing a property or a portfolio — is the way AI is transforming treasury management. There are several core areas where AI is reshaping treasury management specifically for the real estate ecosystem, including:
- Predictive cash flow forecasting: Algorithms analyze historical billing cycles and tenant payment histories to build cash-flow forecasts that dynamically adapt to seasonal rent patterns and unexpected capital expenditure shocks.
- Automated reconciliation: Systems automatically match complex, high-volume bank transactions with internal property management software, virtually eliminating manual data entry errors.
- Virtual account management: AI optimizes how multi-property operators segment their cash balances, providing real-time liquidity monitoring that flags unusual cash positions before they trigger an overdraft or a liquidity crunch.
- Intelligent receivables: The system can automatically identify and clear mismatched payments or partial rent checks, determining which tenant account they belong to without requiring human intervention.
On the fraud and security side, banks are moving away from rigid, rule-based fraud monitoring. Now they’re deploying machine learning (ML) systems that build a unique behavioral baseline for each customer or account, and then flag activity that deviates from it. This strategy reduces the false positives that used to bog down compliance teams and slow legitimate transactions. For a commercial real estate sponsor, critical, time-sensitive funds — including wire transfers tied to a closing — are less likely to be arbitrarily frozen in an automated compliance loop.
The bigger picture
Zoom out, and the common thread connecting valuation, underwriting, treasury, and fraud detection is that banks are becoming faster, leaner, and more data-dense.
This evolution is happening for reasons that extend beyond the borders of commercial real estate. Driving it? Fundamental corporate metrics: efficiency ratios, talent competition, regulatory pressure and the universal forces reshaping any regulated industry adopting new technology at scale.
The CRE angle is, however, worth acknowledging. Since much of banking risk runs through CRE loans, a significant portion of this AI investment is landing on real estate desks first. For developers, investors, and brokers, understanding these internal banking mechanics is key to knowing how your next deal may be analyzed, priced, and approved.
Are you a commercial real estate investor or seeking a specific property to meet your company’s needs? We invite you to talk to the professionals at CREA United, an organization of CRE professionals from over 65 firms representing all disciplines within the CRE industry, from brokers to subcontractors, financial services to security systems, interior designers to architects, movers to IT, and more.
