Beyond the Easy Wins: How AI is finding the next layer of portfolio savings

The first wave of portfolio savings is usually the most visible. Underused buildings. Overlapping leases. Locations that should have been closed two cycles ago. For most occupiers, these are the savings that surface quickly in any portfolio review — and they account for the bulk of what gets reported back to the business.

The harder question is what comes next.

Once the obvious consolidation is done and the visible inefficiencies are out, most portfolios still hold value that's harder to see. Patterns that connect operational performance, lease structures, capital plans and financial signals in ways no single report will show. Decisions that compound across several parts of the business, but don't sit cleanly in any one team's view.

This is where dashboards stop being enough.

From Easy Wins to Hidden Ones: A Recent Case

A global financial and professional services client recently completed a series of mergers and acquisitions that added more than 10 million sq ft to their portfolio. They came to CBRE with a clear brief: identify where space could be optimized, and where occupancy costs could come down.

The first year of the engagement delivered roughly $70 million in realized savings. We used CBRE Vantage® Analytics with Ellis AI to identify consolidation opportunities across the integrated portfolio — downsizing in some markets, relocating in others, closing the locations that no longer earned their place. The work spanned 35+ markets across 30 countries.

That was the first wave. A strong result by any measure.

Then the question got harder. With the obvious consolidation done, where was the next layer?

Where Next Action Engine Comes In

CBRE's Next Action Engine, part of Vantage Analytics, was deployed to take the work further. Running with Ellis AI, it has the same data that fed the first wave — facilities management, lease administration, project management, occupancy and capital planning — and surfaces patterns no single report would reveal.

The output is recommended action, ranked by impact and risk. So far, Next Action Engine has identified an additional $15 million in savings opportunities currently in flight for this client. Some involve capital planning decisions that read differently when held against lease expiry data. Others involve operational signals that, taken together, point to interventions no single team would have flagged on its own.

The actions themselves look like the kind of work CBRE has always done with occupiers. What's different is what triggered them — patterns the team could not have reasonably surfaced by reading each report in isolation.

Why This Applies at Any Scale

This is a large portfolio, and the absolute numbers reflect that. The underlying pattern is independent of size.

Most occupiers already have a strong sense of where the first wave of savings will come from. They know which buildings are underused, which leases are overlapping, which locations have outlived their purpose. That institutional knowledge is valuable — it's what makes the first wave move quickly. The mechanics are well understood: review utilization, consolidate, relocate where the lease economics support it, close what isn't earning its place. CBRE does this work, at scale, every day.

The harder question is what comes after — once the visible inefficiencies are out and the picture looks well-shaped on paper. That's where most real estate teams hit a flat curve. The data exists. The capacity to read it across systems doesn't.

This is the gap AI is starting to close. Tools like Next Action Engine connect data that used to sit in separate reports, then translate the resulting patterns into a recommended next step. The result: the second wave of value becomes accessible to occupiers at any scale, not only portfolios of this size.

At CBRE, this is where the work is moving — toward shorter routes between data and decision, and recommended actions reaching the people who need them without weeks of analysis in between.