High performance has a human requirement

Image created using ChatGPT and Figma by Jemma Frost, 2026

Most organizations don’t decide to stop treating people as their most important variable. It happens gradually, through a thousand reasonable-seeming tradeoffs, until one day the annual engagement survey is the primary way anyone “listens” and nobody can quite remember when that became normal.

The last two years of AI adoption have accelerated that drift, mostly without anyone choosing it. Most rollouts have been framed as efficiency plays: automate the workflow, compress the headcount, measure the productivity gain. That framing is seductive because it’s legible, and it quietly trains an organization to see its people as a cost to optimize rather than a source of intelligence. The tools get the investment and the attention. The humans expected to adopt, trust, and improve them get a change-management meeting.

This is worth naming not as an indictment but as a design problem, and a newly urgent one. Because the same shift that makes AI valuable (work getting more complex, interconnected, and fast-moving) is exactly what raises the premium on the human capabilities no tool replaces: judgment, candor, the willingness to flag what’s broken and try something unproven. The connection between how people experience work and how organizations perform isn’t a matter of philosophy. It’s been measured, replicated, and documented, and it points to a clear set of things leaders can actually do, most of which look less like management and more like design.

If you’re a leader reading this, you probably don’t need convincing that something’s off. You can feel it in the meetings that used to have energy, in the people who’ve gone quiet, in your own depleted capacity to be the manager you want to be. The harder question isn’t whether the human part is fraying. It’s how you’re supposed to tend to it while absorbing more change, more tools, and more uncertainty than any training prepared you for. That’s the real problem, and it’s the one worth solving, because the alternative (waiting for conditions to calm down) is a bet that they will, and they won’t.

Why human-centered culture creates high performance

The evidence here is unusually consistent. When people feel seen, trusted, and safe enough to speak honestly at work, they contribute more, stay longer, and solve harder problems. Safety lowers the cost of telling the truth, and most performance problems are truth problems.

When people feel safe, they surface problems early instead of hiding them until they’re expensive. They offer half-formed ideas instead of sitting on them and they spend energy on work instead of on self-protection. Disengagement isn’t laziness. It’s what happens when the effort of navigating the environment starts outweighing the effort of the job.

You don’t have to take my word for it. The numbers speak for themselves:

• Employees working in human-centric models, defined by flexible work, intentional collaboration, and empathy-based management, are 3.8x more likely to be high performers, 3.2x more likely to stay, and 3.1x less likely to experience burnout. (Gartner, 2022)
• Employees who were given input into their company’s work design were 2.5x more likely to achieve high performance than those who weren’t. (Gartner, 2022 Fortune)
• Employees who receive high-quality recognition are 45% less likely to leave within two years, based on longitudinal tracking of nearly 3,500 employees from 2022 to 2024. (Gallup & Workhuman, 2024 Gallup)

Set against that, the baseline is sobering. Gallup’s 2026 State of the Global Workplace report puts U.S. engagement at just 31%, an 11-year low, which means roughly 69% of employees are disengaged. 50% of U.S. respondents reported experiencing daily stress and 19% reported daily loneliness. Manager engagement itself fell from 27% to 22%, and managers report higher stress than the people they manage.

For human-centered organizations, that gap is the opportunity, because engagement and feelings of belonging are within a leader’s control.

What human-centered culture is

It isn’t a benefits package, a rewards program, or a Slack channel called #wins. It’s an operating philosophy with three commitments:

1. The people doing the work are viewed and leveraged as the best source of intelligence about what is and isn’t working
2. Decisions get made with attention to how they land and not just what they’re meant to accomplish
3. Psychological safety is a condition work is designed for

None of this is soft. High empathy is a prerequisite for high performance. The research on this is not new or controversial. What’s changed is that the cost of ignoring it has gotten a lot steeper. The uncertainty most industries are now facing has raised the baseline of stress people carry into the work, which makes the safety to speak honestly harder to sustain and more valuable when you have it.

Get Emily Stuart’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

When work was more predictable and roles were more siloed, good process could compensate for a poor culture. Those days are mostly over. The problems organizations are trying to solve now are too complex and too interconnected. They require people willing to flag what’s broken, offer half-formed ideas, and work across lines of authority. That only happens when it feels safe to try, and when people haven’t already stopped caring enough to bother. Fear and quiet resignation produce the same silence, for different reasons.

Warning signs that your company is drifting

Most leaders don’t think they’re running a dehumanizing organization. Very few (hopefully none?) set out to create one. But culture is less about intention and more about the accumulation of small decisions over time. Here are some signs worth paying attention to:

• You only hear from people when something goes wrong. If the dominant flow of information is upward-when-bad and downward-when-good, you don’t have a listening culture. You have a reporting culture. Those are not the same thing.
• Your engagement scores are fine, but something feels off. “Fine” engagement scores at an annual cadence tells you almost nothing about the day-to-day reality. If the hallway conversations, the exit interviews, and the things said in confidence don’t match the survey, trust the conversations. Survey data is a lagging indicator, and by the time a number moves, the thing it measured happened months ago. Worse, most surveys are built to produce a score, not the truth, so people give you the polished version that protects everyone involved rather than the one you could actually act on.
• New ideas die in meetings. If you watch people mentally check out when someone proposes something different, or if the response to every suggestion is a list of reasons it won’t work, that’s a culture that has learned risk aversion is safer than contribution.
• Your best people are leaving. We’ve all heard that people don’t leave bad companies, they leave bad managers. In practice they leave both, and usually for a quieter reason than either: the work around the work became harder than the work itself.
• People are doing the job, not the work. There’s a difference between executing a task list and genuinely caring about the outcome. When culture erodes, you often still get the former. You lose the latter, which is exactly where the discretionary effort, the creative problem-solving, and the institutional loyalty live.

Why this matters more as work changes, not less

You may be thinking, “This all sounds great, but if AI is reshaping work anyway, why invest in understanding how it happens today? Why study workflows, roles, and frustrations that might not exist in a year?”

It’s a fair question, but it falls into the trap we all keep falling into in this age of AI. It’s not about what’s changing. It’s about the people.

The main thing you’d be studying isn’t the workflow. It’s the people, and they’re the part that carries over. AI changes tasks; it doesn’t change the need for judgment, trust, and someone willing to say “this isn’t working” before it’s expensive. Those capacities are the least disposable thing you have, and they’re exactly what a human-centered culture practice builds. More than that: you can’t design a transition you don’t understand. “Everything will be different” isn’t a reason to skip the research. It’s the reason the research matters, because you’re now steering people through more change, not less, and the teams that find the new shape of the work first will be the ones safe enough to experiment, honest enough to tell you what’s breaking, and bought in enough to figure out how to fix it. That doesn’t come from a tool. It comes from a culture you build before you need it.

How to start turning it around

No organization is too far gone, but meaningful change doesn’t come from a new platform or process. It comes from applying a discipline you probably already have to a place you’ve never pointed it: the people doing the work. Here are some strategies to try.

• Build trust before you need it. Trust isn’t declared, it’s accumulated, and it’s built or spent in the ordinary moments nobody logs: whether you do what you said you’d do, whether credit lands where the work happened, whether the hard conversation gets had directly instead of routed around. People calibrate constantly on the gap between what you say you value and what you actually reward. You can’t shortcut this with a values deck or an offsite. You build it one reliable interaction at a time, and you protect it by noticing when you’ve spent it and repairing quickly.
• Give people enough of each other to trust. Trust between teammates runs on the same fuel as trust in a leader: knowing the person on the other end. Distributed work, back-to-back meetings, and constant reorgs quietly starve teams of the low-stakes contact where that knowing forms, the hallway version of finding out how someone thinks, what they’re good at, what they’re dealing with. When people don’t actually know each other, every disagreement reads as a threat and every handoff carries suspicion. When they do, they extend each other the benefit of the doubt that lets real collaboration happen. You can’t mandate that closeness, but you can design for it: put people on shared problems instead of parallel tasks, build in the working sessions and rituals where they actually see how each other think, and stop treating the moments that let people connect as the first thing to cut when the calendar gets tight.
• Create the conditions for honest signal. Every design process is only as good as the feedback flowing into it, and fear is the thing that corrupts the data. When someone raises a concern, curiosity has to arrive before defensiveness, because one person shut down in public can quiet a room for months. You’re not being nice. You’re protecting your research from faulty data.
• Understand the experience before you redesign it. When you restructure a team or roll out a tool, start with the people it lands on: what their day actually looks like, where the friction is, what they’re working around that nobody documented. This is where AI adoption most often goes wrong, a tool gets chosen for its capabilities and dropped onto a team whose real experience nobody understood, adoption stalls, and everyone blames the people rather than the rollout. The people doing the work know things the people designing it can’t, which is exactly why you learn from them before you change things on them.
• Prototype in the open, and say what you don’t know. Cultures that demand certainty from leaders get leaders who perform it, and performed certainty produces bad decisions that are then too expensive to walk back. The alternative is to treat a change the way you’d treat a new experience you’re designing: run it small, name it as a test, watch what actually happens, and adjust based on what you see. The thing that prevents chaos here isn’t secrecy, it’s framing. People panic when change feels unbounded and unspoken; they steady when you tell them exactly what you’re trying, why, and when they’ll hear more. “We’re piloting this with two teams for six weeks and we’ll share what we learn” is not vagueness. It’s a more honest kind of certainty than pretending you have the answer, and it gives people something firm to hold onto while the details are still moving. Share back what you find, so transparency becomes a rhythm people can count on rather than a one-time reveal.

What this actually asks

You can invest in every system, tool, and framework available. Their return still depends on whether the people inside them trust each other enough to use them well.

What that takes isn’t softer management or better intentions. It’s a design process, pointed at work itself. The same rigor you’d bring to a product (understand the people who use it, watch where it breaks, test changes against reality instead of assumption, iterate) applied to how work gets structured, how decisions get made, how change gets introduced. It’s the discipline most of us already use on customers and rarely turn on our own teams.

So the human experience of work becomes something you design on purpose, rather than something that happens to your team while you’re heads-down on everything else.

The payoff isn’t only performance. People can feel the difference between being handled as a line in a resourcing plan and being treated like they know something worth knowing.

The technology will keep evolving. The tools will keep changing. The human part has always been the variable that decides what the rest of it is worth, and it’s the one part no reorg or platform will design for you. Build it well and the whole thing gets easier to lead, and easier to live with.