AI pilots surge in CRE, but payoffs lag as leaders pull ahead, JLL says

AI pilots surge in CRE, but results lag. Here's how to turn experiments into ROI

JLL's latest report confirms what many owners, occupiers, and contractors feel on the ground: AI pilots are everywhere, measurable payoffs are not. Pilots jumped from 5% in 2023 to 92% this year, but end benefits trail because teams lack the talent, data foundation, and change muscle to scale.

The takeaway is clear: start small, earn quick wins, and build the plumbing while you go. Companies with solid data, clear processes, and experienced teams are pulling ahead. Everyone else is stuck in proof-of-concept purgatory.

Key stats from JLL

• 92% of companies are running or planning AI pilots (up from 5% in 2023).
• ~27 use cases identified; organizations run ~5 pilots on average.
• Budget pressure hits ~65% of teams, pushing focus to high-impact areas.
• Only 33% of the workforce feels adequately trained on AI.
• 81% report at least three existing systems not delivering expected results.

Where AI is being tested (and where value is most visible)

• Simple, lower-risk starters: lease abstraction, employee experience, location and fit-out strategy, construction planning support.
• High-impact priorities under budget pressure: portfolio optimization, energy management, data workflow automation.

These are sensible bets. Lease abstraction and energy optimization are easy to score and easy to measure. Portfolio optimization and data workflows take more wiring but can move the needle fast once foundations are in place.

The leapfrog myth

AI won't skip steps for you. JLL's data shows the gap is widening: tech leaders are compounding gains, while others stall. If your data is scattered, systems overlap, and processes aren't standardized, pilots stay stuck as demos.

Why pilots stall

• Data sprawl: duplicate tools, dormant systems, no single source of truth.
• Weak change management: no owner, no training plan, no process updates.
• Limited talent: not enough product-minded leaders, data engineers, or analysts.
• Security hurdles: inconsistent access controls and unclear governance.
• Timing: tech rollouts out of sync with budgeting and decision cycles.

A practical playbook for CRE teams

• Pick one use case with clean data and a clear metric. Examples: energy spend per asset, work-order cycle time, days-to-close for lease abstractions.
• Stand up a small cross-functional squad (asset/operations lead, data/IT, finance, vendor). Give them a 60-90 day target and a weekly cadence.
• Measure from day one: baseline, pilot metric, post-metric. Decide to scale, iterate, or kill.
• Rationalize tools as you go: eliminate overlaps, revive what's useful, retire what isn't. Don't add another platform without a decommission plan.
• Stabilize data pipelines: define schemas, lineage, and quality checks for leases, work orders, energy, projects, and portfolio data.
• Bake in change: update SOPs, permissions, and training the moment a workflow changes. No training, no value.

Quick wins you can stand up this quarter

• Energy optimization: dynamic schedules and anomaly detection; quantify savings against weather and occupancy.
• Lease abstraction: automated extraction with human review; track accuracy, cycle time, and legal effort saved.
• Data workflows: auto-clean incoming vendor files; standardize IDs across CMMS, BMS, leasing, and finance.
• Portfolio agility: scenario modeling for disposals, renewals, and reconfigurations with shared assumptions.

Talent and training: the multiplier

Only a third of employees feel ready for AI. That's the bottleneck. If you're investing in data plumbing and new tools, invest equally in the people who will use them - asset managers, facility leads, project teams, and analysts.

If you need structured options to upskill roles across your portfolio and project teams, explore focused programs here: AI courses by job.

Governance and timing

Executives report the best time to change systems is during other major business changes. If that window is open, use it. If not, don't wait. Align AI rollouts to budgeting and approval cycles, and set clear guardrails for data and security.

What to do next

• Choose one quick-win pilot and one foundational fix (for example, energy optimization plus data standardization for leases).
• Define a 90-day scorecard: baseline, target, owners, and weekly checkpoints.
• Create a minimum training plan for the pilot team and end users.
• Set rules for tool sprawl: add one, retire one. Document integrations.
• Prep your scale path: if the pilot works, what assets, regions, or workflows roll in next?

The bottom line

There's no going back. Waiting for a second mover edge risks being left behind while competitors learn faster. Start small, prove value, and build the foundations in parallel. That's how pilots become outcomes.

For context on one enterprise approach mentioned in the report, see JLL Falcon.