2027 predictions: The world of work gets a reality check

In September, a recruiter’s TikTok about iced coffee at job interviews sparked a viral debate. Her point, as she’s since clarified, wasn’t that an iced latte decides who gets hired, but that professionalism still matters in 2026. Since then, the debate (covered by BBC, Bloomberg, SBS News, and others) has continued. The iced latte became a quick, visible proxy for something much harder to measure in a 30-minute conversation: a candidate's readiness, focus, and respect for the opportunity. When what matters is complex to assess, we lean on what’s right in front of us.

That reliance on visible proxies over actual substance is precisely why the corporate world is headed for a reckoning in 2027.

Over the past two years, organizational leaders have been urged to transform their companies around generative AI at breakneck speed or risk irrelevance. Under intense pressure from boards, markets, and shareholders, leaders did what any responsible executive would do. They took decisive action and projected the confidence their teams needed. But in the gold-rush velocity of early AI adoption, the only metrics available were naturally surface-level: software deployment rates, token consumption, and fast restructuring. These visible signals became the enterprise equivalent of the iced coffee. They’re easy-to-see stand-ins for a transformation whose real value was far harder to measure.

In the wake of this rapid transformation, some of the companies that swapped people for AI are hitting Ctrl+Z⁠ (opens in a new tab) and rehiring. The C-suite has a revolving door, with fractional leaders sometimes coming and going before results even come in. Non-HR executives are making workforce decisions, forcing everyone to speak a little HR, which some do more fluently than others. And as AI churns out content faster than anyone can review it, true expertise is no longer knowing the answer, but rather determining how that knowledge can translate into impact. 

The first phase of AI adoption required speed and bold experimentation; the next phase will require precision, patience, and systemic integration. 

As the AI environment evolves, organizations are left with the hard questions: What did we actually get for all those tokens? Who’s accountable when the AI messes up? What is HR for? What even is a job anymore? And what do humans bring that machines simply can’t?

The six predictions that follow are our best answers, compiled from the minds of 34 people scientists, who have met directly with 1162 companies and 177 CPOs in 2026. These ideas were cross-checked against Culture Amp benchmarks spanning over 4000 organizations and 2.6 million employees. 

Prediction 1: Last call in the AI open bar

The era of AI experimentation is ending, and leaders are now expected to show value, not just adoption.

In 2026, "tokenmaxxing" entered the workplace lexicon⁠ (opens in a new tab). At Meta, engineers reportedly competed on a leaderboard to see who could burn through the most AI tokens. Usage was the easiest metric to watch, so it's understandable that companies counted it as impact.

Soon, Meta rolled that approach back⁠ (opens in a new tab). Tokenmaxxing, as we all know, is very expensive. The AI bills are increasing, and it is becoming a harder sell to label AI-spend as R&D. Many AI providers have moved enterprise pricing to usage-based billing, and some analysts argue today's prices are subsidized while the big labs grow ahead of their IPOs. 

In 2027, we expect to see the tokenmaxxing trend morph into its opposite: valuemaxxing. The new goal will be to get the most value from the fewest tokens. 

Therein lies the true challenge for leaders. Valuemaxxing requires a clear strategy. Unfortunately, our new benchmark data shows that a much larger share of employees say they are actively encouraged to explore and experiment with AI (85%) than say their leaders have connected that use to company goals (58%). The plot thickens when we break those averages down by leadership level.

While C-suite leaders are the most likely to use AI habitually⁠ (opens in a new tab), the most senior HR leaders are the least convinced that AI will improve the perception of HR⁠ (opens in a new tab). Why the skepticism? Because "turned-on" AI isn't delivering ROI. In PwC’s 2026 CEO Survey⁠ (opens in a new tab), 56% of executives reported neither increased revenue nor lowered costs from their AI investment.

However, 12% reported major gains in both. The difference? Those high-performing CEOs were two to three times more likely to have embedded AI deeply across the business. 

In a new film, relationship expert Esther Perel and Culture Amp’s founding CEO, Didier Elzinga, discussed organizing humans around AI. Esther argued that the connection between AI and humans at work is “the opposite of polarization.”

In 2027, the leader's job will shift from directing AI experimentation to AI impact with purpose. Leaders will be charged with answering three fundamental questions:

1. Where should employees spend their tokens to drive maximum impact?

2. What should employees do with the capacity created when tasks are automated?

3. How are our people shaping the AI, and how is the AI shaping our people?

Along the way, organizations are going to be forced to contend with what we’re predicting to be the biggest challenge of 2027: accountability. 

Prediction 2: Say hello to “AI ate my homework”

As pressure for AI results grows, accountability will devolve into a circle of finger-pointing between people, managers, and algorithms.

When an employee uses AI to generate a deliverable and the output is dead wrong, who answers for it? The employee who trusted the output? The manager who didn't catch the errors? The prompt engineer? The trainer? The AI vendor? At worst, we’ll see the blame game afoot. At best, confusion about who bears responsibility when AI-assisted work fails will weaken teams. 

It’s like the old excuse,  “The dog ate my homework.” At work, AI has made the blame game a whole lot messier, and we predict it will come to a head in 2027. 

And folks, accountability was already difficult for organizations long before AI took hold. According to our benchmark data, the share of employees who agree their organization manages underperformance has stayed remarkably steady since 2021, moving just 1% point between then (49%) and now. And even Peak Performance cultures⁠ (opens in a new tab) struggle to transparently address employees who underdeliver (59%).

Recent IBM research⁠ (opens in a new tab) affirms that accountability at work is becoming a defining CHRO issue. They found that 71% of CHROs see employees’ ability to supervise, validate, and override AI as essential. But they also found that two in five CHROs believe that employees do not feel safe challenging or overriding AI outputs. 

Colby Nesbitt, Head of Talent Experience Research at Netflix, contends that AI is ushering in even bigger existential questions about how we define labor in the first place.

Ultimately, we expect accountability in 2027 will come down to one non-negotiable principle: You own the output, no matter who, or what, wrote the first draft. What you turn in is what gets graded. But defining "your work" when AI drafted it is a performance question, and performance questions land with HR. 

Across recent CHRO role changes⁠ (opens in a new tab) and HR guidance⁠ (opens in a new tab), people leaders are arguing that HR must help define where AI can act, where human judgment remains, and who is responsible when systems get it wrong. What a company decides will set the tone for accountability in 2027.

Prediction 3: CPOs face a fork in the road

In 2027, Chief People Officers will choose to be the executors or architects of change, and the road that feels safe may be the riskier one.

2026 has been a hard year to be a CPO. Change keeps arriving faster than the strategy to make sense of it, and for many HR leaders the sensible response has been to keep their heads down, automate the admin, and avoid making waves. One HR professional in our AI in HR survey shared their fear that boards might decide “the whole [HR] value prop can just be delivered by AI.” It’s hard to take big swings when you suspect your function is the next line item to be cut. This leaves many HR professionals feeling like change is happening to HR, not through it.

In 2027, we expect two paths to become clear. Some CPOs will become executors: keeping the function running as it always has, using AI just enough to make it faster and leaner, and taking direction from the business on what change they need to lead through next. Others will become architects: designing the change instead of waiting to receive it, and redefining what a people function is for. Both are understandable. Only one feels safe, and that feeling is the trap.

AI is already pulling HR in different directions. In our AI in HR survey⁠ (opens in a new tab), the share of responses saying HR should own AI strategy rose 10% points in a year, while confidence that AI will improve the perceived value of HR expertise fell 9% points in the same time frame.

More want the wheel, but fewer expect a personal return (and we’re still talking about less than half on a good day). Not everyone gets to choose, either. SHRM’s 2026 survey⁠ (opens in a new tab) of 1,908 HR professionals found that 52% of organizations don’t involve HR in AI strategy, directly or through cross-functional work. For those CPOs, the fork starts with earning a seat.

So what happens down each road? The executor CPO gets a faster, leaner function, but the risk is that strategy drifts sideways. AI, workforce, and org design decisions get made by the CFO, COO, or CEO; HR is left executing choices it didn’t shape; and other functions start doing people work themselves. In some companies, the people function ends up reporting to another function instead of leading the people agenda. That’s where the sidewalk ends: not replacing HR overnight, but moving the strategic part of the job down the hall.

The architect CPO takes on the harder questions first. At the outset, they ask, “How does the organization and culture need to change in a world of AI?” and then, “How does the HR function need to align to that change?” Take competency models as an example: The executor CPO updates them to account for changing needs. The architect CPO questions if competency models are the right approach when roles are changing so rapidly. What each needs most is a way to demonstrate to others that their actions are effective. According to our previous research, getting the first core HR process transformed by AI is a key ingredient in moving AI skeptics in your organization into believers⁠ (opens in a new tab). The payoff is a function that pulls other teams toward it, not around it.

For CPOs ready to architect change, Carmen Amara, Yelp's Chief People Officer, recommends starting with a clear picture of how AI adoption is evolving across the organization.

Neither road is easy, and architecting transformation will feel like the riskier bet. But standing still is a choice too, and it’s the one with the bigger downside. The transformation road runs straight through the work itself, starting with a question that’s harder than it sounds: What is a job, anyway?

Prediction 4: Job crafting is back, baby!

Redesigning work will require leaders who set the metrics and employees who shape the tasks.

“So what do you do?” That classic dinner party question is getting harder to answer,  because all of the ways we categorize work suddenly seem outdated. Even three years ago, Deloitte found that only 19% of executives and 23% of workers said work is best structured through jobs⁠ (opens in a new tab). And other options, like skills, are shifting too quickly to keep up, with LinkedIn projecting that 70% of the skills used in most jobs will change by 2030⁠ (opens in a new tab). We’re [this] close to HR being asked to answer that big existential question: “What even is a job?” 

With all that ambiguity swirling around, we predict that job crafting will become a norm again in 2027, but within a new set of constraints. And the companies that do it best will take a dual approach to redesigning roles. 

They’ll start with the human, recognizing that people will, by nature, move toward what they like in a role and away from what they don’t. Job crafting⁠ (opens in a new tab), a term coined by Amy Wrzesniewski⁠ (opens in a new tab) and Jane Dutton⁠ (opens in a new tab), looks at work design with the person in mind. More specifically, it’s the idea that employees actively reshape their work in three ways: tasks (what you do), relationships (who you do it with), and cognition (how you think about it).

Nick Lichtenberg⁠ (opens in a new tab), Fortune Business Editor⁠ (opens in a new tab), says millennials are living out that famous Strokes title, "Is This It?" Glassdoor research⁠ (opens in a new tab) suggests they have hit mid-career and, in many cases, peak earnings potential.

With AI in the mix, task crafting has become much more accessible. When employees want to offload irksome tasks, they can more readily pass them off to an agent. The people who most effectively determine which tasks should be automated and which should be augmented are the ones doing the work. And the good news is employees expect to be brought along, with 4 out of 5 believing they will have a say in how AI is rolled out⁠ (opens in a new tab) in their organization.

But crucially, leaders have a role to play too. The dual approach requires a top-down process where leaders evangelize the vision and set the core metrics for each role. Without that clarity, employees may craft their roles to focus on less critical parts. This is probable, given that we’re seeing heavy AI users feel much more productive but with the same perception of workload as light users. 

Since 42% are unclear on how AI connects to the company strategy, it’s anyone’s guess where that additional work capacity is going. So why not solely take a top-down approach? Because the further away you are from a role, the higher the likelihood you may misunderstand what a task requires or automate away the work that feels most meaningful to the employee. 

This isn't the same job crafting of the tech boom, though. That period of abundant budgets and ballooning headcounts has passed. Today, we suspect job crafting will often happen in scarcity, so practicing it will be less about designing a dream job and more about deciding what to do with the capacity automation frees up. Employees can't reimagine a role without knowing what it's for. So the leader's job is to set the guardrails. Define the vision and KPIs for each role. Be explicit about what's open to change and what isn't. 

If you’re taking a bottom-up and top-down approach, where do they meet? The team.

Prediction 5: The team becomes the unit of culture

Culture is the system. The team is where it’s practiced, and where leaders will aim in 2027.

Leaders who are invested in culture likely have a few ways to understand it (we offer some at Culture Amp, btw). They have annual engagement surveys, pulse checks, retention dashboards, and more. But they tell us that the hard part is acting on the problems their culture measurement tools diagnose. 

Culture is the system, but teams are its micro-cultures: the places where behaviors harden into norms. Because leaders in 2026 are under pressure to show results⁠ (opens in a new tab), and because constant change can make the organization feel more abstract⁠ (opens in a new tab) to its employees, the team becomes the unit people can see. That, and new AI tools are now able to pull the signals teams generate into a single view, so a leader can see what’s happening at the team level alongside those broader org-wide culture analyses. 

We predict that in 2027, leaders will start treating the team as the unit of culture, using their periodic, organization-wide diagnostic data to act on culture team by team, all year long.

Our own Regional Director of People Science in EMEA, Arne Sjostrom, says this isn’t the first time leaders are recognizing the power of teams.

Our data suggests⁠ (opens in a new tab) it’s a sound bet. 82% of employees agree “In my team, we consistently meet or exceed the high goals we set,” versus 64% on the organization-level version of the same question. Part of that is proximity, since people rate higher what is closer to them. But it is also where behavior change effort is most influential. We see the same pattern when it comes to sustainable performance behaviors.

Team-level gelling is linked to performance, too. Employees who have good working relationships with people on their team are 39% more likely to be rated high performers. It’s an association, but it demonstrates just how important team relationships are to performance. And to us, it tracks. Ask people what they like best about work and, so often, the answer is the people they work with. That’s good news, leaders. You don’t have to move an entire organization to move culture. You can start with the relationships employees already value most, in the teams they already show up for.

In a year dominated by what machines can do, the most reachable culture lever may also be the most human one. People like people. And people like human-powered work.

Prediction 6: Human-powered rises up as a value proposition 

Facing a barrage of AI-created work and content, employees and consumers alike will yearn for what is indelibly made by humans. 

Every company adopting AI is making a decision, whether they realize it or not. AI can be a way to do the same work with fewer employees, or a way to do more innovative, value-creating work with the employees you already have. It’s understandable that the first path is tempting, since the savings show up in next quarter’s numbers. But cost-cutting has a cost of its own, paid in customer, employee, and candidate trust. We predict that in 2027, “human-powered” will become a claim that companies make and customers expect.

Let’s start with customers. We’ve all been there. You’re scrolling through LinkedIn, and you spot it: a post that actually looks like it was written by a human. Your heart alights! And research shows you’re not alone. In two experiments⁠ (opens in a new tab) with about 300 adults, researchers randomly labeled the same AI-generated paintings “human-created” or “AI-created.” People rated the “human” ones as more profound and worth more. A separate survey⁠ (opens in a new tab) found that given identical items at the same price, 68% chose the one labeled “made by humans” and 10% chose “made with AI.” And 54% said they’d pay more for the human-made product. 

Importantly, these are hypothetical choices rather than actual purchases, but as AI-generated content floods everything, “human-powered” starts to look like the new “organic.”

Candidates are sending a similar signal. In Greenhouse’s research, 38% of US candidates noted they have dropped out of a hiring process because it included an AI interview. We’re not saying organizations should be Luddites. Notably, in that same survey, more US candidates wanted more AI with stronger human oversight (22%) than wanted less AI (19%). So the trick is picking where to deploy AI and where to keep the human element. And we predict great talent will want to work for organizations that strategically deploy humans. 

Leaders, meanwhile, seem to be heading the other way. In Culture Amp’s global benchmarks, we’ve seen the percentage of employees agreeing to “Leaders demonstrate that people are important to the company’s success” falling from 76% in 2022 to 69% in 2026. 

Of course, this is employee perception, not leader intent, and layoffs and restructuring are highly plausible contributors, so we can’t pin it all on AI. But the gap is hard to ignore, and we predict that 2027 is the year when a company’s approach to AI could become a real value proposition for those striking the right balance.

But as this trend continues, there are also risks worth naming. First, there is potential for “human-powered” to become the new greenwashing, and customers and employees will spot the gap between the claim and the experience. Second, if human attention becomes a premium tier, it could turn these products into luxury goods, available only to an elite wealthy class.

So what should leaders do? Focus on “moments that matter.” Map every place AI touches your customer and employee experience. Decide where the experience must be human, and why. Then measure it: Ask employees which parts of their job they wish were automated, and ask customers whether they reached a person when it mattered.

All together now

Maybe the iced coffee debate got one thing right. Work is full of visible proxies. It is too easy to keep mistaking what is easy to see for what matters.

In 2027, the best organizations will have the keen eye to spot the difference. They will measure outcomes, name ownership, redesign roles in collaboration with the people doing them, invest in teams, and protect the moments where human judgment and connection make the difference.

Employees already expect a say in how AI enters their work. They also know the future is unwritten. Leaders don’t need to pretend they have every answer, nor do they need to have a perfect plan in place. 

We can let AI write the first draft and make suggestions. We did build it after all. But humans must decide what is worth writing in the first place.

So bring on the agents, copilots, automations, and AI outputs. And yes, bring the iced coffee, too. Just don’t let the cup become the focal point. In 2027, let’s grade the work, grow the people, and build workplaces where the future feels less like something that’s happening to us and more like something we’re building together.