At a glance
- By 2026, 83% of respondents had moved past test stages in at least one area and 40% said they are scaling or embedding AI across all activity areas.
- For allocators weighing a manager's AI claims, the checkable evidence sits in asset and property management, the survey's lowest-ranked priority and its strongest source of measurable efficiency gains.
- JLL frames the constraint as integration, not access: large language models have made deployment easier, and the work now is embedding tools where they create lasting value.
Investment outperformance and capability building in capital deployment are real estate investors' top AI priorities at 49% each, according to JLL's 2026 Global Real Estate Investor AI Survey, first reported by IREI. Both also sit among the areas delivering relatively limited results.
By 2026, 83% of respondents had moved past test stages in at least one area and 40% said they are scaling or embedding AI across all activity areas. In 2025, 88% were only testing or planning AI programs. The median investor runs seven use cases across five business functions. Eighty percent report tangible benefits, but 43% say they still struggle to achieve their goals. Only 15% have built custom agentic workflows across every deployment area.
For allocators weighing a manager's AI claims, the checkable evidence sits in asset and property management, the survey's lowest-ranked priority and its strongest source of measurable efficiency gains. Capital deployment, ranked first at 49%, is where results run thinner.
JLL frames the constraint as integration, not access: large language models have made deployment easier, and the work now is embedding tools where they create lasting value. The gap between 40% embedding AI across all activity areas and 15% with custom agentic workflows everywhere is where integration remains incomplete.
Deep AI integration remains rare among real estate investors
Share of investors at each depth of AI adoption, 2026 survey
Show as table
| Label | Value (%) |
|---|---|
| Beyond testing, any area | 83% |
| Embedding across all areas | 40% |
| Custom agentic workflows, all areas | 15% |
Reported by the Sectors desk from IREI. Published . Corrections