powering productive workplaces
Front page
NewsWorkforce Analytics1h · 09:01 BST · 10 min read

The Q4 Space Decision: What Occupiers Lock In This Month

CBRE’s new occupier research and October’s workplace technology releases have given facilities teams more to work with as 2027 budgets are signed. The big decisions are still about space, meeting rooms and trustworthy data. AI planning tools could help teams get them right.

Work analytics
Work analytics

I think we all need to give up on the idea that a brilliant dashboard can save anyone from an expensive lease mistake. October’s the month to ask what we’re actually signing for in 2027: how much space, which meeting rooms need investment, and which data source everyone will trust when the numbers disagree.

And vendors are certainly giving us more ways to answer. HubStar launched AI-powered Dynamic Planning on September 30, OfficeSpace introduced Studio on September 17, and Robin followed with a Claude connector on October 6. VergeSense has added 12-month shortage and surplus forecasts, too.

We’re moving from software that tells you how busy Tuesday was to tools that can model which floors you'll need next year. There’s genuine potential to shorten a process that used to take weeks of spreadsheets and competing assumptions.

Most of the strongest savings examples still come from occupancy measurement rather than these newer AI planning features, though. I’d want that distinction written into the business case before anyone signs another contract.

Are Companies Expanding or Cutting Office Space Going into 2027?

I talked about this in my last office utilization roundup, but the numbers are getting harder to square with the idea that everyone’s about to downsize. CBRE’s 2026 Americas survey found 23% say AI already influences space planning, with another 30% expecting an impact within two years. That’s perception, not proof they’re all using AI planning software.

In Asia Pacific, CBRE surveyed more than 650 corporate real estate executives in September; 50% expect to expand within three years. I’d be wary of signing a smaller lease assuming automation or remote working will keep emptying offices.

Space availability complicates things. JLL’s October 13 US report puts leasing growth at 6.4% over 12 months, with 9.6 million square feet of inventory disappearing in Q3. Its October 16 Central London update recorded 3.1 million square feet of Q3 take-up, the strongest third quarter since 2019, and 5% quarterly growth in active demand. Get your requirements wrong now, and replacing that space could be expensive.

CBRE’s September analysis expects U.S. office-using employment to grow 0.9% annually over the next five years. Meanwhile, half of Americas occupiers think AI will increase demand for flexible, reconfigurable offices. Neither figure tells you how many desks you’ll need to buy.

Before committing to more space for 2027, I’d want to know whether existing rooms are the wrong size, how teams might change, and what reconfiguration would cost compared with another lease.

What Office Utilization and RTO Assumptions Should Shape 2027 Budgets?

I wouldn't let finance sign off a 2027 workplace budget built around one utilization percentage. It looks great until everyone turns up on Tuesday because someone promised cookies, and half the team can't find somewhere to sit.

CBRE recorded average global office utilization of 53% in 2025, up from 38% in 2024, across portfolios covering 303 million square feet. Peak utilization reached 80%. Assigned occupancy was 111%, meaning more people were allocated space than there were seats available. It doesn’t mean they all turned up together, as we explained in our earlier meeting-room ROI article.

CBRE’s 2026 Americas survey adds another complication. The share of employers requiring at least three office days rose from 78% in 2025 to 89%. Actual attendance averaged 2.9 days, though, against an expectation of 3.2. That 0.3-day difference belongs in your budget calculations.

Johnson Controls shared an interesting July example: a workplace leader expected another mandatory office day to lift utilization by 20%. It rose just 2%. It's an unnamed, vendor-reported case, but I wouldn't fancy explaining that forecasting error to finance.

In the UK, Acas opened a flexible-working guidance consultation on October 1, ahead of anticipated statutory changes in autumn 2027. It's no new right to work from home, or a forecast of attendance, but another reason to test policy scenarios.

For RTO planning, what happens when policy meets an ordinary week? Are departments coming in together? Could staggering schedules relieve Tuesday's pressure?

Get demand by building, floor, team, and room type. Test busiest recurring days and actual team overlap, not just the weekly average.

Which AI Planning Tools Could Change Q4 Space Decisions?

At least the planning tools are catching up. HubStar's September 30 Dynamic Planning release uses The planning tools are catching up with what companies need right now. HubStar’s September 30 Dynamic Planning release uses workforce figures, attendance policies, and occupancy data to model lease and growth scenarios. Fit&Fill and AutoFit suggest where teams could sit. HubStar estimates it can eliminate 90% of manual planning work. I’d want that measured during an actual move before putting it in a budget.

Robin took a different route on October 6, adding a Claude connector and conversational Space Planning. Facilities teams can test layouts and leave proposed changes in draft until someone approves them. Sensible, when an AI suggestion could become an expensive furniture move.

OfficeSpace Studio lets teams edit imported plans and publish bookable layouts. Its CTO, Kiley Reynolds, described the old process:

“A space planner used to send a request to a CAD team and wait days to move twenty desks.”

OfficeSpace says Studio now takes minutes. More interestingly, published changes can update connected booking, capacity and utilization data, with versioning to track revisions.

Spacewell’s September 29 Predictive Planning release, developed with VergeSense, models headcount growth, floor closures and attendance-policy changes, although exports and imports are still involved. VergeSense’s October 1 Meridian update adds 12-month assessments of possible space shortages and surpluses. It can recommend converting underused spaces, then rerun forecasts to show the projected impact. What happens if the sensor data is incomplete, or the proposed change costs more than it saves?

That’s what I’d ask every vendor: once the new layout looks right, do bookings and occupancy reports reflect it, or is someone still updating three systems? Finance might find that answer more useful than the AI demo.

Can Workforce Data Support Reliable 2027 Space Forecasts?

Well, that depends. Even the cleverest planning tool will struggle if HR expects 200 new employees while facilities budgets for 50. You can ask AI to rearrange seats fifty ways; it still needs to know who’s coming.

Pinterest workplace planning lead Brian Huber told HubStar in September:

“Headcount growth has a much smaller impact on office utilization than actual changes in attendance.”

Pinterest uses the 80th percentile of its three busiest office days rather than a one-off peak. Huber also explained that HR’s headcount, finance’s approved hires and employees’ likely cities can tell different stories. Plan another floor using the wrong version, and you’ll get an expensive surprise.

Gartner reported on October 7 that only 38% of CHROs and CIOs share an understanding of AI’s impact on future work. That’s a worrying starting point for predicting who’ll need desks.

JLL’s 2026 benchmark found utilization data informs 90% of planning decisions, yet only 7% rate their occupancy-data capabilities excellent. CBRE found 48% primarily use Excel, while 86% integrate HR feeds, 59% audit space data manually, and 15% have no governance. Dashboards aren’t the problem; trustworthy inputs are.

OfficeSpace’s February report analyzed records from 954 organizations covering 116 million square feet, including over a million scheduled moves. That’s platform data, not a global survey, but a reminder that floor plans get stale quickly.

VergeSense’s October update accepts multiple sensor brands alongside badge, Wi-Fi, and booking records. Helpful for mixed estates, provided discrepancies can be explained.

I’d want every forecast to show whether demand changed because of attendance, hiring assumptions, or outdated floor plans. Otherwise, it’s another argument about whose number is right.

Can Workplace Analytics Cut the Cost of Running the Office?

Only if companies start asking the right questions. For instance, if half a floor sits empty every Friday, why are we still paying to heat and clean it as though everyone’s there? I’d put that question to facilities before approving another analytics subscription.

Johnson Controls’ 2026 research identifies energy optimization as a leading use of AI in facilities management. There’s a more concrete example of what occupancy data can achieve, too.

VergeSense reports that a global investment bank connected occupancy data to Switch Automation at a New York office, adjusting cleaning and HVAC to actual building use. The reported savings reached $2 million annually on cleaning, alongside daily energy savings of up to 30%.

Those are vendor-reported results, so I’d want to see the baseline before promising similar returns elsewhere. Ventilation and comfort requirements still apply, even on quieter days.

Still, getting a facilities bill to reflect how a building is actually used is a result finance can measure.

How Do You Build a Defensible 2027 Workplace Budget?

If a vendor says its AI knows exactly how much space you’ll need next year, I’d ask how well it would’ve predicted last year’s attendance. JLL’s 2026 benchmark found 46% of organizations cite cost as a barrier to occupancy-planning AI, while 45% struggle with system compatibility. A tidy one-building pilot is one thing. Twelve sites, aging badge readers, and consultants still fixing imports next spring are quite another.

I’d keep five questions ready for the sales team:

  1. “Our AI predicts your space requirements.” How accurate was its last forecast, by site and room type?

  2. “Everything’s in one dashboard.” Which integrations and data-cleaning costs are included in the price?

  3. “Our agent optimizes your office.” Who implements its recommendations, and what will those changes cost?

  4. “AI cuts real estate costs.” Which actual expense falls after licensing, integration, and support fees?

  5. “Planning updates automatically.” Can our team rerun scenarios with revised hiring plans without another consulting bill?

I’d approve a limited deployment with a named owner, agreed accuracy targets, and a review date. If the forecast misses those targets, finance should be able to stop spending rather than inherit another subscription nobody wants to challenge.

The Final Test for Your 2027 Workplace Budget

For all the excitement around AI workplace planning, I'd still want to know whether the technology could've helped us make a better decision last year. Test its predictions against historical attendance it hasn't already seen, across different locations and space types. Then see how well it handles low-, central-, and high-demand scenarios when the assumptions change.

I'd also want to know what happens after the forecast. Can facilities turn a recommended layout into something people can actually use, without weeks of manual updates or another consulting bill?

For me, that's what makes a 2027 workplace budget defensible. Finance needs evidence that the forecasts are reliable, the recommendations are practical, and the promised savings can actually be measured. I'd take that over a dazzling five-year AI roadmap any day.

FAQs

What should finance ask before approving workplace technology?

Ask what problem disappears if the spend works. Then ask how anyone will prove it disappeared. If nobody can name the metric before purchase, the business case is already weak. “Better employee experience” won’t survive a budget review unless somebody has attached it to room failures, support load, space pressure, or another cost finance can see.

Can we forecast office demand using booking data alone?

You can try, although a room booked for eight hours might only be used for forty minutes. That's a fairly big difference when you're deciding whether another floor is necessary. Without some way of checking actual attendance, you're making expensive decisions based on people's calendars.

Is it too late to test AI planning software before 2027?

Depends when your lease decision is due. If the landlord needs an answer in November and the vendor needs three months to prepare your data, the timing's already a problem. Ask how soon the platform can produce something useful using your actual building records.

What happens if an AI workplace forecast turns out to be wrong?

You’ll want to know who can change the assumptions and whether every correction comes with another consulting bill. If the forecasts keep missing, how much trouble will it be to take your historical data to another provider? Those answers belong in the purchasing conversation.

rate this story
helps rank stories across uc today
The discussion0 takes · attributed & checked

Does this reflect your experience?

opening the room…
Read nextordered by techtelligence · every pick explained
picked for this story

The 53% Utilization Trap: Redefining Meeting Room ROI

10 Oct 2026
picked for this storyWorkday Rising Verdict: The Latest Promises Tested17 Oct 2026picked for this storyAI Expiry Week: GPT-5.5 Exposes the AI Lifecycle Problem15 Oct 2026