Optimism and uncertainty battled to start the year for commercial real estate, with the momentum of late 2025 colliding with the reality of a fraught early 2026.
Sitting at the junction of the two are underwriters, the group of analysts, attorneys and loan officers whose job it is to evaluate risk, project cash flow and perform financial modeling.
Like many in commercial real estate, underwriters are trying to find the best ways to incorporate artificial intelligence into their workflows. But the nature of their work means that no matter how much they rely on machines, it still has to be a human making the calls.
“I wouldn’t give AI $20M to invest,” said Robb Gilman, an accounting and audit partner at Anchin.
Deal flow picked up to start 2026, with the Mortgage Bankers Association predicting total commercial mortgage origination volume will hit $806B this year, up from $633.7B in 2025.
Although that estimate came out three weeks before the conflict in the Middle East that has sent shockwaves through global markets, most remain optimistic that a relatively short-lived conflagration will keep CRE on the recovery track where it began the year.
Underwriters and financing experts see companies racing to gain the speed and competitive advantages expected by those who are embracing it. As AI underwriting evolves out of its infancy, the industry may begin to see a mix of “horror stories and success stories,” Crexi Vice President of Product Growth Adam Siegel said.
But especially in an era of increasing uncertainty and emerging data around macro risks, having a longtime expert reviewing or finalizing decisions seems to be the preferred model.
“It’s about having a throat to choke,” JLL Global Head of Strategy and Quality Management Becci Curry said. “Somebody wants someone, not just an AI, to blame if something goes wrong.”
There is certainly a new sense of caution around investments in Middle Eastern property, Curry said. But a traditional measurement of geopolitical risk, the 10-year Treasury index, saw a spike in February, a peak in late March and a downward slope ever since.
Broadly speaking, there hasn’t been a massive disturbance of deal flow, despite caution and concern over the potential of war to cause more global disruption, said Michael Riccio, CBRE co-head of national production for capital markets debt and equity finance.
Despite an inflation surge over energy prices, the interest rate remains transactional, he said, and CBRE is on track to see a 16% year-over-year increase in sales activity, with similar action on the debt side.
Underwriting has evolved in recent decades as the real estate industry has become more fixated on shorter-term investments, Riccio said.
Once a strictly long-term asset, real estate has become a traded commodity, he said. Short-term holds of three to five years are more the norm as the industry’s biggest fish raise massive funds. It is all about chasing the highs and lows of the market amid the volatility, which changes the strategy and risk appetite.
In light of this shift, AI’s ability to more quickly evaluate and source opportunities makes it a valuable investment. Much of underwriting focuses on the terms and clauses of complex and lengthy leases, Siegel said. AI offers help in more quickly deciphering these documents, as well as environmental reports and offering memorandums.
The increasing speed made possible by AI is shifting the expectations of younger professionals.
“AI is changing what we expect junior analysts to be good at faster,” Riccio said. “Underwriting has always been about judgment. What AI does is take friction out of the process — cleaning data, running scenarios, pressure‑testing assumptions — so analysts spend less time building spreadsheets and more time understanding real risk.”
The industry-standard tools for underwriting and analysis include Argus and customized spreadsheets. Siegel said both have their downsides. Argus can be expensive and feel “archaic” to use, while spreadsheets can break down if a small part of the formula is disrupted.
New AI tools can factor in fresh datasets or rethink traditional assumptions, according to Mike Broder, CEO and co-founder of RCKRBX, which provides demand-side data for the multifamily market.
His platform uses geographically customized datasets around the local workforce, especially the propensity to work remote or hybrid schedules, and helps developers adjust the unit layout and amenity profile of multifamily projects to obtain more favorable underwriting.
Broder’s company will survey thousands of people in the region and map out a demand profile of jobs and salaries over the next 36 months, then extrapolate assumptions about unit mixes and building layouts. Retail projects can use foot traffic data to strengthen the case for shopping center investments.
In the case of offices, it is possible to run more analyses of the tenant mix and evaluate their vulnerability to AI cutting jobs, therefore reducing occupancy.
Larger real estate companies even find value in curating unique data sources to aid in underwriting. Avison Young has its own Office Busyness Index, meant to be a more accurate and actionable measure of workplace usage.
This kind of data can help improve a deal’s risk profile, said Connor Burke, vice president of debt and equity finance for Avison Young.
“These investment opportunities go up to a small pool of investment committee members that ultimately make a decision,” Burke said. “Very compelling data points go a long way in helping move forward with an opportunity.”
Siegel said the increased speed and accuracy of AI-based underwriting will level the playing field and allow more players to move forward more rapidly with more deals.
It could also make things easier by widening how and where deals are sourced, since AI underwriting will more quickly find similar opportunities. But there is still broad belief that amid the improving science of underwriting, there is an art to finding and identifying opportunity.
“It’s like the robot surgeon,” Siegel said. “You’re not going to want to be the first person to do the robot surgeon by itself. You want a doctor there. It’s the same thing here. Understand that this is the future, but we may not be quite there yet.”
