Designing for (and with) Anxiety

Zoom background tidy. Bluetooth headphones charged and connected. Quick intros. Then ten minutes in, it was clear that I had prepared for a different version of this interview.

The recruiter had walked me through what to expect from the hiring manager, and I built for exactly that. I armed myself with a canonical study doc: product style guide notes, voice and tone, competitors, market CAGR, case study details mapped to likely questions. I rehearsed the arcs. I timed the walkthrough. I staged answers to the follow-ups I expected. It was a thorough piece of preparation with one structural flaw, and I know this flaw well, because finding it in products is my job. I only designed the happy path. The interviewer was friendly, and their questions were adjacent to everything I had prepped. But adjacent is often where the happy path ends, and the edge cases live, and I showed up to an edge case with no recovery flow.

After I stumbled through my answer to the first question, a familiar tinnitus pulse rang in my ear. My heart rate climbed. I tried to ingest the question cleanly, but I kept locking into the case study flow I had practiced, and I never fully found my groove again. Then the interview was over. I put my head on the desk for a few seconds, then went back to the freelance project I had open, dejected.

Later, I thought about what had actually gone wrong, because the freeze-up made no sense on paper. Pivots and thinking on my feet are hallmarks of my design leadership career. And when that first question landed, and it wasn't an invitation to walk through a case study, an internal alarm went off, the same one I have trusted for many years, and it said what it always says: "something isn't right." Most of my career, I have acted on that signal on behalf of products, teams, and the business. This time the signal was pointed at me, and it was carrying more load than one interview should: a year of searching for full-time work, a news cycle full of AI layoffs and restructuring, a referral from an ex-co-worker whose name was on the line next to mine for this interview. So I paused and hoped the alarm would subside. When it did not, I pushed the prepared material harder. I heard my own confidence drain out mid-sentence. I bombed an interview for a job I was qualified for and wanted. And the part I could not shake afterward was this: the thing that bombed it was not a weakness. It was the same trait that built my career.

It took me a while to understand how both of those things can be true. To explain it, I have to start with the strange relationship my profession, Design, has with anxiety. We do not just experience it. We work in it. It is our raw material.

Almost no one opens a work application inspired and excited. Data shows that the apps that make people happiest are the ones they spend the least time with. They arrive carrying a task and a low hum of worry. Am I doing this right? Did that save? Is this the button that deletes everything? I have a problem, and I am hoping, without much evidence, that this screen solves it.

Most of what we call good product design is partially anxiety management work. A good jobs-to-be-done definition should capture more than just the task: how well the user understands their own problem, how stressed they are, and what is at stake if they can't finish. An undo command exists so that a mistake is not a catastrophe. Delightful success indicators elevate users' mood and provide confidence that they are on the right track. A confirmation state exists so that "did it work" has an answer. An empty state exists so that a blank screen reads as a beginning instead of an error. Autosave, inline validation, progress indicators, recovery flows. None of it makes the user's alarm go away, and it should not try. It does something better: it answers the question "what do I do next?" It wires the alarm to an action.

A smoke detector on its own is helpful, but it's often just noise. It assumes you know what the next steps are, how serious the problem is, and whose job it is to respond. What makes it a system is what it is wired to: the thing that wakes the building, calls for help, starts the response. The difference between software that feels calm and software that feels hostile is rarely the pretty UI. More often, it is the comprehensive UX: ensuring every alarm the user might feel has somewhere to go. There are not many alarms on the happy path. Feeling good is good, until it is not and you do not know what to do next.

Product designers are professionally fluent in this. Sometimes the alarm is a failed upload; other times, a password reset or an MCP server fail. We spend our whole careers giving other people's alarms somewhere to go.

The leadership version runs on the same fuel, with one difference. This time, you are the alarm.

The best design leaders I have worked with share the trait that never appears in a job description: they see failure early. Three months out, they can feel that the roadmap will not hold. They notice the handoff that is going to crack, the headcount plan that does not match the strategy, the stakeholder whose silence is not agreement. That was me for most of my career. At one company, the alarm fired early in a flagship launch: the scope that product and engineering had committed to did not match what the teams could actually execute. Surfacing it was uncomfortable; shipping it as planned would have been worse. At another, the alarm went off on our own process, not a product: we were prioritizing by output instead of impact, and saying so out loud is how the team ended up adopting Shape Up and changing what we said yes to. I saw where things were going to go poorly, and looked to surface it and adjust. Sometimes it worked. Sometimes it did not. But the pattern held: detect early, discuss it, and change the plan while changing it is still cheap.

Psychologists once ran an experiment where a room slowly filled with smoke from what looked like a broken computer. The people who scored high on anxiety noticed first, and in the follow-up study they were also the most effective at warning everyone else. Researchers call it sentinel behavior: some people are wired to catch ambiguous signs of trouble and raise the alarm for the group. Every team I have ever worked on had a sentinel. It was usually me.

A team learns what is safe to say by watching what its leader says out loud.

I used to describe that as diligence: treating every dollar, every hour, every strategy as if the company were mine. The honest word is anxiety. Some of what drove me was care for the team and the product, and I stand by that part. But a lot of it was simpler and less flattering. I was wired to make sure nothing went wrong on my watch: the stakes were too high. The pressure to fix a problem before it became an issue was not a methodology. It was a fear of being the person standing next to the failure. It was a survival mechanism, and ingrained survival mechanisms do not come with an off switch.

The vigilance does not clock out when you do. The same wiring that catches a failing handoff at work will wake you at two in the morning to re-inspect a Slack conversation that went fine. An alarm with no off switch scans whatever is in front of it, and on a quiet Sunday the only thing in front of it is your own life. The career gets the benefit. The person pays the bill.

Pointed at the work, with the authority to act on it, that alarm is close to a superpower. That is the sentence I built a career on, and I have spent years tuning the alarm to my advantage. I still believe it. I just learned the hard way that it is only half a sentence.

Which brings me back to the interview. The detector fired correctly and on time. Once when the recruiter's email said to have "a case study prepared to answer these questions," and I wondered whether I should ask for clarity. And again in the room, when the off-script questions started, and the alarm said: "this is not the conversation you prepared for." Every other time in my career, the thing at risk was a product, a plan, a launch. This time the thing at risk was me. And that changed everything about what I did with the signal. Both times, nothing.

The move was sitting right there. It takes a few seconds. "Great question. To be honest, I prepped for a different conversation. Can you repeat the question and give me a second to recalibrate?" I have said versions of that sentence in rooms for years on behalf of products and strategies. I have said it in kickoffs, in design reviews, in war rooms. "Do we have shared clarity?" and "what would make this fail?" unlock alignment and trust. Naming the meeting you are actually in is the most basic move a design leader owns.

If a product is in trouble, I say so early and fix it in the open. When the interview was in trouble, I hid it and tried to fix it silently, in my head, between sentences. Part of that was pride. Most of it was a belief that has served me my whole career: that the answer is always within reach if I keep working the problem. That belief is why you hire someone like me. It is also what kept me quiet, because saying "I prepped for a different conversation" out loud would have meant admitting, for a moment, that I did not have the answer.

That is the real shape of anxiety in this profession, and it completes the sentence I built a career on: detection wired to action is an instrument. Detection with the wire cut is an alarm nobody answers, and everything the fire takes is yours. The alarm in that interview did not fail me. It worked perfectly. Nobody answered it. The nobody was me.

Most people hide anxiety at work. Many of those same people have a Thursday block between 2 and 3 pm marked "Busy" that is actually a therapist appointment. So I will say the quiet part out loud: anxiety is not a flaw to be coached out of leaders, or anyone. Hypervigilance just wears different name tags at work: the naysayer, the ocean-boiler, the something-is-off-er. Strip the name tag off and you usually find someone ten steps ahead, trying to make the thing succeed: pushing for clarity, getting teams to work better together, making sure the software is good and that people will actually buy it. John Cutler wrote recently about this wiring under its kindest name, the person who thinks too much, and his conclusion matches mine: it is a mismatch with the room, not a defect in the person.

We, the sentinels, push for research. We get prototypes in front of real customers. We poke at a PRD's soft spots and call the CX team to hear a customer describe the problem in their own words. And when the wiring is given authority, the moves get bigger: it kills the roadmap item nobody else wants to question, restaffs a team before the crack becomes a crater, walks a launch back a week out because shipping would cost more than the delay. Different name tags, same instrument.

The work, for people built like me, is not to feel less. It is one move, and it is the move I did not make. When the alarm fires on you, say it out loud, in the room, while adjusting is still cheap. The words are situational. "This is not the conversation I prepped for." "I am not confident in this plan, and I want to be." "I need a minute." The move is always the same: name your own state before it names you. It is the undo command for a person, and it is not a dodge of accountability but the way you keep it: you cannot own an outcome you have quietly lost the means to deliver. Every designer reading this has built that mechanism for a stranger. Fewer of us have ever built it for ourselves. And for leaders, the move is bigger than any one meeting: a team learns what is safe to say by watching what its leader says out loud. Go first, and you have not just wired your own alarm. You have wired the room's.

I have made the ask at full scale, all the way up to a CEO: "I understand what you expect from design and product. Here is what the team needs to be accountable for it." The answer that came back was silence. Silence is an answer too. The ask does not guarantee the answer you hoped for. It guarantees you know where you stand, and that what you do next is a decision instead of a drift.

I do not get that interview back. What I get is the read. The alarm never cost me that room. Treating myself as the one product that had to fail silently did.

If any of this sounds familiar, you are probably a sentinel too. There are more of us in this profession than the job descriptions admit, doing our best work because of the alarm and paying for it in private. The alarm is not the thing to fix. The silence is. This essay is me breaking mine. What would you say out loud, in the moment it matters, if you treated yourself like something worth protecting? Start with what you need. Name it to yourself, then ask for it out loud. Once you have it, it is all on us. Unless we never ask.