AI and the future of work are now the same conversation. Enterprise AI adoption sits around 80 percent, according to the 2026 State of the Workplace report, and time spent inside AI tools grew eightfold in two years. Yet the practical questions are still open. How long should the workweek be? Why do AI investments go unused? What happens to focus? And what does it all mean for office space?
Corporate real estate, HR and workplace leaders face these questions from different sides. The answers below draw on practitioners, researchers and Kadence conversations from 2024 to 2026, organized around the decisions leaders actually have to make.
AI and the future of work: what is actually changing?
AI is changing three things at once: which tasks people do, who does the work, and how fast the business expects answers. It automates routine work, makes AI skill a basic job expectation and raises pressure on planning. On the current evidence, it is not removing the need for a workplace.
Start with tasks. In March 2025, Kadence’s Jamie Addis, arguing that AI is making the five-day workweek obsolete, cited a McKinsey & Company report finding that AI could automate up to 60-70% of employee tasks in data processing and administrative roles. He also cited Accenture research that organizations integrating AI into their operations report 2.5 times higher revenue growth and 2.4 times greater productivity.
Then who does the work. In April 2025, Shopify CEO Tobi Lütke told staff that using AI effectively was a fundamental expectation of everyone at the company and that hiring would be “AI-first.” As Kadence co-founder Dan Bladen explained in why Shopify proves AI-first is the new remote-first, this repeated a pattern. In May 2020, Shopify declared that “office centricity is over” and rebuilt around async work and intentional in-person gatherings. Bladen’s view is that the real risk is bolting AI onto rigid, presence-based ways of working: “You can’t be AI-first if your operating model is 2013.”
By 2026 the picture was more nuanced. At IFMA in April 2026, Brian Elliott and Dan Bladen discussed why workplace strategy can’t stay static in the age of AI. They agreed that new tools raise what organizations expect from their teams, and that in practice this often creates more complexity before it creates simplicity.
Will AI make the five-day workweek obsolete?
Some leaders think so. Atom Bank CEO Mark Mullen argues that as AI takes on repetitive and administrative work, a five-day week will stop making sense. His bank introduced a four-day work week in 2021, cutting hours from 37 to 34 with no salary reduction and no drop in performance.
Jamie Addis made the case in March 2025 that AI is making the traditional five-day workweek obsolete. At Atom Bank, employees reported feeling more engaged and recruitment became easier after the change. Mullen put it bluntly: “I think a four-day week isn’t progressive, it’s bloody logical.” He expects the trend to continue, and says he cannot understand how “a five-day week, which is a pre-Victorian invention, still works in the digital age.”
Addis’s broader argument is that the real productivity gains come from companies that trust their teams, give them autonomy and use AI to create smarter ways of working, not from forcing people back into offices. The shorter week, in this view, is a strategic response to AI rather than an employee perk.
What is a compressed work week, and how does it relate to AI adoption?
A compressed or shorter work week gives people back working days. Workforce strategist Gillian Brookes uses it as an incentive for AI adoption: teams that find at least 10 percent productivity gains through AI earn a day off every two weeks. The reward gives people a personal stake in using AI well.
In June 2026, Brookes set out the AI incentive playbook for using shorter work weeks to accelerate AI adoption. She calls the core problem the kitchen appliance problem: a new gadget sits on the bench looking impressive while everyone keeps cooking the way they always have. Research she cites shows 80% of employees now use AI tools, but adoption remains shallow. People avoid deeper use because they sense that mastery might make their role redundant.
Her evidence comes from a five-month pilot of a shorter work schedule, with one day off every two weeks. People started with personal habits such as time-blocking, then aligned focus hours with colleagues. Three months in, teams began working with adjacent teams and a culture of continuous improvement took root. Brookes says she saw a workforce mature faster than anything she had witnessed in 15 years of organizational development.
Her three-step framework:
- Set clear expectations. Make AI adoption a stated priority and track AI usage alongside traditional productivity metrics.
- Remove fear through commitment. State that the purpose is not redundancies, and prove it by sharing the productivity benefits through flexible work.
- Run a structured pilot. Set a six-month pilot with a clear value exchange: at least 10 percent productivity gains in return for a day off every two weeks.
The capacity math matters to business leaders. If employees receive a 10 percent productivity gift while the organization gains 25 percent overall through AI, there is 15 percent additional capacity to deploy. One of her manufacturing clients used it to reimagine quality control. Brookes is also running pilots in manufacturing and construction, where the reward becomes flexible scheduling, shift pattern optimization or additional rest periods.
Incentives are one condition. Understanding the work is the other. In a November 2025 conversation, riding into 2026 with Dan Bladen, Brian Elliott described Zapier’s code red moment in 2023, when it realized it had to map its processes before AI could improve anything. AI creates value when companies understand what good looks like, where human judgment adds value and where automation removes unnecessary friction.
How do you protect focus when AI speeds up work?
Design for human capacity, not just throughput. Workplace strategist Omar Ramirez points to data showing average focus sessions have fallen to just over thirteen minutes as AI use grows. His answer is screen-free analog spaces, deliberate positive friction and measures of human outcomes alongside utilization.
In June 2026, Ramirez described the analog advantage of designing for human capacity in an AI-saturated workplace. The 2026 State of the Workplace report he cites analyzed more than 443 million hours of work activity. Collaboration activity is up 34 percent. Research from UC Berkeley and Harvard Business Review points the same way: when AI accelerates work without a counterbalance, people produce more while thinking less clearly. Ramirez calls this the AI productivity paradox.
His observation is counterintuitive. The most technologically advanced organizations are investing in analog spaces: rooms with no screens and no devices, just whiteboards, paper and conversation. These rooms quickly become the most-booked spaces in the building. Ramirez says organizations using Kadence to track occupancy often see analog spaces fill faster than screen-heavy conference rooms that cost far more to build.
He also argues for positive friction. A coffee station with a three-minute natural wait gives a real conversation time to start. Some firms are reintroducing slower elevators in specific zones, because the extra thirty seconds turn into cross-floor collaboration. The practical rule is to match the mode to the work: strategic planning, creative exploration and trust-building in analog settings; status updates and task tracking in digital ones; data analysis digital, with AI in the loop.
Ramirez recommends four steps. Audit the floor plate and score every zone by tech intensity. Design space typologies from what the audit shows. Prototype one space for ninety days before scaling. Then measure human outcomes such as belonging scores, unplanned-to-planned interaction ratios, dwell time and retention signals. Utilization tells you how often a space is used. It cannot tell you what was created there.
Is hybrid work the right foundation for AI?
Yes, according to several Kadence conversations. Hybrid already runs on outcomes, async collaboration and team agreements, which are the foundations AI needs. Brian Elliott reports that hybrid has stabilized at around three days a week, and that clear alignment on goals matters far more than the number of office days.
In September 2024, Kadence’s Future of Work Strategist Dave Cairns launched Beyond Hybrid, Kadence’s series on the next chapter of work. Back from the Running Remote conference in Lisbon, he told a panel: “This is a hybrid moment. Period.” He noted that the stock performance of in-office companies lagged their fully flexible peers by 16% between 2020-22. His advice to large enterprises was incremental change, moving from top-down mandates to team-led agreements.
The return-to-office push that opened 2025 tested that view. Speaking with Dan Bladen in November 2025, Elliott said attendance had increased, but not to the level leaders commanded:
“Three days a week seems to be the average now. The number of people moving to five is less and less every month. What we are seeing is stabilization rather than escalation.”
Companies like OpenAI, which set a clear Monday through Wednesday cadence years earlier, saw the smoothest adoption. The conversation’s deeper point is that hybrid struggles are very often alignment struggles. In a conversation looking ahead to 2026, Elliott described a CEO running six business units across twelve countries who pushed back on a simple office-day mandate because the real challenge was clarity:
“It was not the process. It was getting senior leadership aligned on how to clearly articulate the most important five goals for the company for the next six months. Alignment is the actual work. Office days are just the surface layer.”
That lesson carries straight into AI adoption. Organizations that measure presence struggle to absorb new tools. Organizations structured around outcomes can change how work gets done without changing what success means.
Will AI replace the office?
No. In an April 2026 conversation at IFMA, Brian Elliott and Kadence co-founder Dan Bladen noted that AI companies such as OpenAI keep growing and taking office space. AI is instead raising scrutiny on cost and making static space planning hard to defend. The leaders handling it well compare scenarios before they commit.
The economics are sharper than before. In November 2025, Elliott described a company paying forty thousand dollars per desk per year for premium office space. Companies are dividing floors into premium client-facing space, flexible collaboration areas and efficient working zones, and right-sizing rather than cutting at random:
“If your max number of people in seats over the last ninety days was thirty, you are getting thirty seats. There is no point in keeping sixty when the data shows you only need half.”
Workplace teams now carry that pressure. Their remit has grown from leases, layouts and facilities to space strategy, utilization analytics, hybrid coordination and reconfiguration planning, at the intersection of real estate, people operations, IT and finance.
Static planning breaks down under these conditions. Headcount shifts faster, attendance stays uneven and business units use space differently. A plan based on average attendance can miss where demand peaks, and a space model can overlook the adjacencies a team needs. For AI space planning, the inputs matter as much as the model. Seat counts, contracts, attendance signals and growth forecasts often sit in different systems, and AI can help bring that picture together, but it cannot compensate for weak inputs.
Bladen shared two examples at IFMA. One customer believed a new floor needed 100 additional desks. After modeling team adjacencies, likely peak demand and overcapacity risk, the real answer was two. Another needed to find 15 desks on a floor in Sydney and used Kadence to identify where that capacity could come from. In his words: “Static planning gives you a snapshot. Scenario planning shows you consequences.” That is the case for scenario planning as a core workplace capability.
The market backdrop raises the stakes. Writing in August 2026, Bladen cited the PwC and Urban Land Institute Emerging Trends in Real Estate report: 49 percent of office leases in place in March 2020, at ten thousand square feet and above, still have not rolled over, and net absorption has been flat or negative for fourteen straight quarters. His conclusion: you cannot lease your way out of that market. You have to optimize what you already hold.
Who owns the workplace demand signal?
The occupier does. Dan Bladen argues that sensors and digital twins describe the building, while the demand signal describes the workforce: bookings, attendance, team patterns and headcount plans. Employees generate that data, so it sits with the occupier and the workplace leaders closest to it, not with the building.
The 2026 PwC and ULI report, in its 47th year, ran a spotlight titled From Proptech to PropOS. It describes a property operating system: AI agents, digital twins and data integrations on top of existing building platforms, connected rather than replaced.
In from PropTech to PropOS: who owns the demand signal, Bladen agreed with that shape but challenged where the intelligence comes from. A sensor can count how many people sat on the third floor on Tuesday. It cannot tell you that two teams should sit together on Tuesdays, or that a business unit is about to add forty people. The report itself shows why behavior matters more than policy: only 72 percent of organizations hit their attendance goals, and just 37 percent took any action to enforce their return-to-office policies.
His summary line: “Supply data without demand data is a dashboard. Demand data with an optimization layer is an operating system.” Consolidation, co-location and lease renewal decisions all need the demand signal plus an engine that turns it into comparable scenarios.
The first step is unglamorous: audit whether your demand data is captured anywhere structured. For most organizations it lives in calendar invites and a hand-updated spreadsheet named FINAL_v7. In Kadence’s approach, WorkOps captures the demand signal as employees coordinate their week, and SpaceOps turns it into portfolio scenarios using constraint-based optimization rather than a generic language model, so every scenario traces back to the organization’s own workforce data.
How should leaders manage change as AI reshapes work?
Involve people early, test ideas in context and facilitate rather than police. Culture change consultant Tara Birch, PTHR founder Perry Timms and smart working expert Andy Lake each argue that lasting change comes from involvement, evidence and trust. The same applies to AI adoption, where fear is the main barrier.
In October 2024, Kadence’s Helen Attia spoke with Tara Birch about her Change Together Cycle for navigating change together. Birch treats change as an ongoing cycle, not a one-off project. Involve employees from the start instead of handing down decisions. Build strong internal communities; she once built a Culture Ambassador Network of 500 volunteers across the globe. Stay flexible, gather feedback and be ready to pivot. She also creates safe spaces where leaders can admit their own fear of change.
Later that month, Perry Timms wrote about moving from thought leadership to sense-making. His example is a CEO who reads that rescue teams perform better when they eat lunch together, then orders everyone back to the office, although the company is full of designers and software developers. The better move is to ask what your own version of that cohesion looks like, then test and learn. As Timms puts it: “Does thought leadership help you make a business-critical decision? Probably not; the data and evidence should do that.” He also reminds leaders that, using estimates that 30-40% of work is office-based, the office debate affects at most 1.36 billion people in a 3.4 billion global workforce.
In April 2024, Attia discussed transforming the public sector with Andy Lake. Their conversation framed smart flexibility as something woven into how an organization operates, built on trust. It starts with leadership, needs buy-in from the executive team to frontline staff, and requires managers to shift from overseeing to facilitating.
Where to start
- Make the deal explicit. Name AI adoption as a priority, rule out redundancy as the goal and share the gains.
- Map workflows before adding tools. Define what good looks like and where human judgment matters.
- Pilot and measure. Track productivity, engagement, retention and how space use changes week by week.
- Protect focus. Give deep work and trust-building a physical home.
- Plan in scenarios. Capture the demand signal and model growth, capacity and policy before committing to space.
AI will keep changing the pace and shape of work. The organizations that benefit will not be the ones with the most tools. They will be the ones that give people a reason to use them, protect the human work that tools cannot do, and base space decisions on evidence.