The case for
Forward-
Deployed AI
in CRE
Published with CRE Daily
Commercial real estate has stopped asking whether AI matters. In a recent Ollo survey of 100+ CRE professionals, 71.5% expect their firm to increase AI spending next year, and almost none expect a cut. But adoption is shallow: most firms use AI as scattered individual habits, not something built into how the business runs. The firms breaking through aren’t buying the most tools — they’re bringing in a partner to build AI directly into their existing workflows. This report calls that model forward-deployed AI, and the data suggests it is where CRE is heading.
How CRE firms are approaching AI adoption
Share of firmsAI Everywhere, Scaled Nowhere
CRE has moved past exploration. Two years ago, 76% of firms told Deloitte they were piloting or researching AI; today 66% use AI tools weekly or daily (First American Data & Analytics / DealGround). But using AI and trusting it are different things — only a sliver hand real decisions to it:
How CRE professionals use AI today
Share of respondentsAdoption also splits sharply by size — from 28.7% at the smallest firms to 52.4% at those with 100–249 employees. That gap points to a widening divide: larger investors and brokerages can absorb the cost and complexity of AI, while smaller firms risk being left behind. The largest players are already pushing past admin work into property operations and higher-value analysis (ULI, Emerging Trends in Real Estate 2026).
What the Survey Found
Ollo surveyed more than 100 CRE professionals, weighted toward decision-makers — 46.5% were owners or founders — with focus areas skewed toward the deal side.
Respondents by primary focus area
Share of respondentsHow they use AI tells the real story. Most rely on it as individuals; only about a fifth have woven it into recurring operations.
Firms’ current use of AI
Share of firmsThe Overlooked Move: A Partner Who Builds It With You
As the survey shows, only 3.3% of firms have hired an outside partner to build and manage AI. That number is low for a simple reason — the option barely existed a year ago. But the ceiling on the alternatives is becoming clear. Almost every firm now has an “AI person,” and they are often the first to say they fell into the role.
I’ll ask about their background and hear, ‘I came in as a broker, started vibe-coding a couple years ago, and now I run all our AI.’ They can do impressive things on their own. But standing up agentic systems that run reliably in production is a different problem.
It is less that these teams are lost than that there are hard limits to what a non-technical user can scale. The underwriting example below is the gap between a clever one-off and a system that runs 500 deals without breaking. As firms hit that ceiling, Han expects the 3.3% to climb fast:
“That expertise usually doesn’t exist inside the institution. Once firms treat this like infrastructure, it gets built by people who do it full time.”
What Forward-Deployed AI Looks Like
The model borrows from technology: instead of selling software and leaving, a team embeds with the client. Ollo is a services business, not a product. Its people sit alongside a CRE team, map how the work is done today, and build AI into that process — the model the analysts already use, the reports the firm already sends. There is no new platform to learn, which removes the change-management barrier that stalls most tools. Engagements start with a free four-week pilot on the client’s most painful workflow, with weekly working sessions and on-site enablement. Ollo’s founders have worked together since 2018 and shipped 25+ products used by 10M+ people.
- 1A free four-week piloton the client’s most painful workflow
- 2Weekly working sessions
- 3On-site enablement
First-Pass Underwriting
A multifamily operator (15,000+ units, $2.5B AUM) had analysts spending 1–4 hours on a first pass for every one of ~500 deals a year — roughly 2,000 hours, mostly on deals it would pass on. Ollo built the first pass into the team’s existing model, deal room to reviewed output. Every figure Ollo validated — taxes, capex, opex — came within 5% of the underwriters’ own numbers.
- 7 minutes with Ollo
- 1 hour, the low end before
- Up to 4 hours, the high end before
First-pass underwriting time per deal. Source: Ollo.
Monthly Reporting
A national asset manager and receiver overseeing 1,000+ properties spent about 300 hours a month on financial reporting — chasing property managers, auditing data, and rebuilding 200+ pages by hand, 16,000+ reports a year. Ollo built two agentic systems, one to collect documents and one to generate reports, mapped to the firm’s own logic. Work that took two weeks now runs in about a day.
Monthly reporting time, before and after. Source: Ollo.
A 2025 JLL survey found 88% of real estate firms are piloting AI, but only 5% have hit all their AI goals. The difference isn’t ambition — it’s whether the AI is wired into where the work actually happens.
1 dot = 1% of real estate firms. Source: JLL, 2025.
The Bottom Line
CRE has decided AI matters. What’s unsettled is how to run it. Building and buying both leave the hardest part — making AI work inside a real operation — to the customer. Forward-deployed AI closes that gap by putting the people who build these systems inside the process itself. Today only 3.3% of firms have taken that route; the ones already seeing measurable ROI suggest the rest won’t wait long.