Corbis Intelligence
Research and Insights
Commercial real estate intelligence grounded in academic research, real economic data, and the agent systems behind Corbis and EQUIRE.
Search articles by title, content, or tags...
Commercial real estate intelligence grounded in academic research, real economic data, and the agent systems behind Corbis and EQUIRE.
Agentic Assets Research Team
Corbis Research
February 24, 2026
8 min read read
In February 2026, commercial real estate services stocks became a public test case for AI anxiety. The market reaction was not subtle. Reporting from The Guardian described sharp declines across CBRE, JLL, Cushman & Wakefield, Savills, IWG, British Land, and Landsec as investors rotated away from labor-intensive businesses they viewed as exposed to AI automation. A later Wall Street Journal piece put the issue more directly: brokerages had to persuade investors that humans still matter in selling, leasing, and valuing buildings.
The selloff was not just a story about software replacing brokers. It was a story about information costs. Brokerage, appraisal, capital markets advisory, and research all bundle public data, private market knowledge, relationships, judgment, and execution. AI attacks the parts of that bundle that are repetitive, document-heavy, or dependent on data aggregation. It does not automatically replace negotiation, local context, capital relationships, or accountability for high-stakes recommendations.
The first margin pressure is likely to appear in the work that precedes the client meeting: compiling comps, summarizing rent rolls, screening tenants, extracting lease clauses, normalizing offering memoranda, and drafting market narratives. Those tasks are not trivial, but they are increasingly tool-addressable. The McKinsey Superagency report, published in January 2025 and still useful for February 2026 readers, found that most companies expected to increase AI investment, while only a small share considered themselves mature in deployment. That gap is exactly where real estate firms sit: high intent, uneven operating maturity.
AI will compress the time required to produce a first analytical draft. That matters because the traditional model often monetizes the hidden labor behind the deck. If a client can get a credible first pass faster, fee pressure follows. But the output still needs evidence discipline. A fast comp table is not the same as a defensible valuation conclusion.
The Financial Stability Board's 2024 report on AI and financial stability gives a useful framework for real estate finance. AI can improve efficiency, but it can also introduce model concentration, data-quality failure, cyber risk, third-party dependency, and misleading confidence. In commercial real estate, those risks show up when every firm draws from similar data vendors, uses similar models, and mistakes automated fluency for market truth.
The best broker or adviser in an AI-enabled market will not win by manually doing every task. They will win by controlling provenance: which sources matter, which claims are current, which comps are stale, which underwriting assumptions are fragile, and which parts of the answer deserve human review. That is a different service model from staff-hours-as-output.
The market was right that AI will pressure high-fee businesses whose value proposition is mostly information packaging. It was too blunt if it treated all CRE advisory work as the same. Standardized tasks are exposed first. Complex capital decisions remain relationship-driven, evidence-heavy, and judgment-dependent.
The durable lesson for owners, lenders, and investment managers is not to ask whether AI replaces the broker. Ask which part of the workflow is being automated, what evidence survives the automation, and who is accountable when the recommendation is wrong. In 2026, that is the new line between a marketing deck and a decision system.
The practical question for owners is not whether a brokerage firm uses AI. It is whether the firm can show how evidence moves through the advisory process. A stronger stack should preserve source dates, separate public market data from relationship knowledge, and flag when a claim comes from a comp set, a broker opinion, or a model. That distinction matters because AI can make weak evidence look polished. The useful workflow is therefore less about replacing brokers and more about making the broker deck testable.
The February selloff also fits a broader labor-market debate. The IMF's work on generative AI and labor exposure argues that advanced economies face high exposure because many cognitive tasks can be complemented or automated. CRE advisory is exactly such a market: the work contains judgment, but it also contains large amounts of repeatable synthesis. The firms that protect margins will be the firms that turn repetitive synthesis into a controlled system.
The selloff was useful because it forced an uncomfortable but productive distinction. AI will commoditize generic information packaging. It will make accountable interpretation more valuable.
Comprehensive analysis of AI adoption in real estate finance, featuring market projections, platform comparisons, and strategic implementation guidance for 2026.
Discover how artificial intelligence is transforming commercial real estate in 2025, from predictive analytics to smart building management, driving 70-95% efficiency gains and creating new competitive advantages for property professionals.
