When people are anxious, they stop reading the way designers imagine. They skim, they jump to the scariest word on the screen, and they fill any gap with the worst possible explanation. A label that looks neat in a design review can feel cold or alarming to someone who is already worried.
I learned this most clearly on Pukaar.ai, a doctor-led app that helps parents understand why their baby is crying. Our core user was a tired parent at 2 a.m. The words on the screen mattered as much as the layout, sometimes more.
This post is my working approach to UX writing for anxious users. It applies to health apps, banking, insurance, legal tools, or any product where people arrive worried and need to leave calmer.
Why labels fail anxious users
Designers love labels. They are short, they fit in a chip, and they make a system feel organised. "Needs Soothing", "High Risk", "Pending Verification". The problem is that a label tells the user what category they fall into, but not what it means for them or what to do next.
For a calm user, a label is a useful summary. For an anxious one, it's a verdict without an explanation. Their brain does the explaining, and it rarely chooses the gentle version.
That is why one of the four design decisions on Pukaar.ai was simply: plain language before the medical label.
The "sentence first" principle
On the Cry Report screen, the wireframe originally led with the category. In the final UI, the category ("Needs Soothing") was demoted below a plain-language explanation. The parent reads a calm, human sentence first, and the label sits underneath as a summary.
The order matters because people read top to bottom and stop early when they're stressed. Whatever comes first sets the emotional tone for everything after it.
A useful test: cover everything on the screen except the first line. If a worried person read only that line, would they know roughly what is going on and feel a bit steadier? If not, rewrite it.
A simple structure for anxious-moment screens
I now use this order on almost any screen that delivers news, results or status:
- What's happening, in one plain sentence.
- Why, with reasoning the user can check.
- What you can do now, as clear, visible actions.
- The label or technical term, for people who want it.
On Pukaar.ai, the second step became its own block called "Why I think that". It reads back the parent's own sleep and feed history and shows that answers come from doctor-verified sources. When people can see the reasoning, they are much more willing to trust the conclusion. I cover that idea more broadly in designing visible trust in fintech.
Practical rules I follow
These are the rules I come back to when writing or reviewing UX copy.
Use the words your users use
Research tells you the vocabulary. On Pukaar.ai, parents in the WhatsApp community asked questions in everyday language, not clinical terms. If your users say "my baby won't stop crying", don't answer with "persistent distress vocalisation". Mirror their words first, then introduce a term if they need it.
Note
Many Indian users mix Hindi and English in a single sentence. If your research shows that, test copy in the mix people actually speak, not in formal Hindi that nobody types at 2 a.m.
One idea per sentence
Short sentences are easier to process when attention is low. If a sentence has an "and" joining two different ideas, split it.
Tip
Read your copy aloud in one breath. If you run out of air before the full stop, the sentence is too long for a tired reader.
Lead with the outcome, not the system
"Your transfer is complete" is better than "Transaction status: processed". The user cares what happened to their money, not what the system calls it.
Say what happens next
Uncertainty is the enemy of calm. Every status message should answer "what now?" On Demerara Bank, a persona we designed for, Shanice, is a nurse who wants certainty that a transfer went through. That is why the transfer flow has a clear form, review and success step, so the user is never left wondering.
Watch out
Don't promise a timeline the system can't keep. "Usually within a few working days" is honest. "In 24 hours" that turns into a week breaks trust faster than saying nothing.
Never hide the escape route
For anxious users, a visible way out is reassuring even if they don't use it. On Pukaar.ai, escalation to a paediatrician is always visible, and a red-flag check on every reply from QARA, the AI companion, can hand over to a consult in under 15 minutes. The copy around that route should be calm and direct, not alarming.
Avoid false cheerfulness
Exclamation marks and "Oops!" work in a game. In a health or money product, they can feel dismissive. Warm is good. Chirpy, at the wrong moment, is not.
Before and after examples
Here are a few rewrites in the style I use. They are illustrative, not copied from any shipped product.
| Before | After | What changed |
|---|---|---|
| "Error 403: Authentication failed." | "We couldn't sign you in. Check your password, or reset it in a minute." | Outcome first, plus a way forward |
| "Result: Discomfort." | "Your baby may be uncomfortable, perhaps from wind or a wet nappy. Here's why we think that, and what you can try." | Sentence before label, reason and action added |
| "Application status: Under review." | "We're reviewing your application. You'll get an update here, usually within a few working days." | Says what happens next and where |
| "Are you sure? Yes / No" | "Send ₹5,000 to Rahul? Send money / Go back" | Specific question, verb labels |
Notice the pattern: plain sentence first, reason where it helps, specific action labels instead of "Yes" and "No".
Writing with research, not guesswork
Good UX copy isn't a final polish step. It comes from the same research that shapes the flows. Real quotes from interviews are often the best source of tone. On Demerara Bank, a customer named Savitri said "I can never find where to start", and another, Andre, said "so much on the screen at once". Those two lines told me the copy had to be fewer words, clearer headings and an obvious first action.
I also use AI tools like Claude and ChatGPT to generate alternative phrasings quickly, then edit hard against the research. AI is good at giving you ten options. Choosing the one that fits a scared parent at 2 a.m. is still a design judgement. I write more about that balance in how I use AI in my UX workflow.
A quick checklist for anxious-moment copy
Before shipping a screen that delivers news or status, I check:
- Does the first line make sense on its own?
- Is the technical label below the explanation, not above it?
- Can the user see why the system reached this conclusion?
- Is there a clear next action, written as a verb?
- Is there a visible way to get human help?
- Would this sentence sound kind if read aloud to a worried friend?
Work with us
Want a team to design it with you? Our UI/UX design agency for startups and SaaS takes products from research to developer handoff.
Key takeaways
- Anxious users skim and assume the worst, so the first line sets the emotional tone.
- Put a plain sentence first and the label second.
- Show the reasoning, then offer clear actions with specific verbs.
- Use the words your users use, gathered from research, not your internal jargon.
- Always keep a calm, visible route to human help.
Words are part of the interface
Typography, spacing and colour help, but in stressful moments it is the words that people remember. If you want to see how plain language shaped a real product, read the Pukaar.ai case study, and pair this post with designing for the worst moment. If your product's copy needs a rethink, drop me an email.