Healthcare · AI product design · 2026

An AI wellbeing companion for Bay Area tech workers caught between burnout and an 8-week therapy waitlist.

MindBridge is a self-directed case study exploring a real, current healthcare gap: the weeks someone spends knowing something is wrong, with nowhere immediate to take it.

Data points sourced from public 2025–2026 reporting on Bay Area mental health access. The product, research quotes, and participants below are illustrative.

My roleLead UX Designer, self-directed
Timeline12 weeks · 2026
PlatformiOS & Mac, full parity
ToolsFigma AI, Claude, Maze
80%
of one Bay Area therapist's caseload now works in or with AI
8 wk
average wait for a first therapy appointment in San Francisco
35k
Bay Area tech layoffs recorded across 2025
62%
of tech professionals report feeling drained at work
The interactive prototype

Meet MindBridge

A working prototype, not a screenshot — take the burnout scan, get a real score, and see how the AI journal and Bay Area resources respond to it. Every screen works identically on both devices.

9:41●●● 100%
MindBridge

Good evening, Maya

Sunday · 14-day streak

Wellbeing
73
+6 this wk
Burnout
— —
Not yet scanned
Streak
14
days
Daily check-in
How present did you feel at work today?
Get started
Take the 3-minute burnout scan to unlock personalized insights and your AI journal.
Dr. Aisha Patel
Burnout · tech anxiety · SF
● Available Tue

Burnout risk scan

8 questions that adapt to your answers. Under 3 minutes.

Question 1 of 8

Your burnout profile

Here's what your 8 answers suggest.

/ 100

Take the scan first

Your journal prompts are shaped by your burnout profile. Complete the 3-minute scan to unlock it.

Bay Area resources

Filtered for tech workers, near you.

Tech therapists
24 verified providers, telehealth ready
Walk-in centers
No appointment needed, Santa Clara Co.
Peer groups
8 weekly meetups across the Bay
Psychiatry
Same-day telehealth available
Need help right now?
Call or text 988 · Text HOME to 741741 · Bay Area Crisis (415) 781-0500
Maya Rodriguez
maya.r@proton.me
Account
Email
maya.r@proton.me
Password
••••••••
Plan
Free
Privacy & data
Hide from workplace integrations
Never visible to employer or HR systems
Share location for matching
Used only to find nearby therapists
AI uses entries to personalize
Never used to train shared models
Export my data
Notifications
Daily check-in reminder
8:30 pm, gentle tone
Streak reminders
We never guilt you for missing a day
Therapist availability alerts
Appearance
Dark mode
Match the case study toggle above
Support
Crisis resources
Help & FAQ
MindBridge is a wellbeing companion, not a substitute for licensed care. Entries are private and never sold.
MindBridge v1.2.0
iPhone — quick check-ins, on the go
MindBridge wellbeing companion
Home

Good evening, Maya

Sunday · 14-day streak · reviewing from your Mac

Wellbeing
73
+6 this wk
Burnout
— —
Not yet scanned
Streak
14
days
Daily check-in
How present did you feel at work today?
Get started
Take the 3-minute burnout scan to unlock personalized insights and your AI journal.
Dr. Aisha Patel
Burnout · tech anxiety · SF
● Available Tue

Burnout risk scan

8 questions that adapt to your answers. Under 3 minutes — works the same on Mac as on iPhone.

Question 1 of 8

Your burnout profile

Here's what your 8 answers suggest.

/ 100

Take the scan first

Your journal prompts are shaped by your burnout profile. Complete the 3-minute scan to unlock it.

Bay Area mental health resources

A bigger screen for a slower, more deliberate search.

Tech therapists
24 verified, telehealth ready
Walk-in centers
No appt needed
Peer groups
8 weekly meetups
Psychiatry
Same-day telehealth
Need help right now?
Call or text 988 · Text HOME to 741741 · Bay Area Crisis (415) 781-0500
Maya Rodriguez
maya.r@proton.me
Account
Email
maya.r@proton.me
Password
••••••••
Plan
Free
Privacy & data
Hide from workplace integrations
Never visible to employer or HR systems
Share location for matching
Used only to find nearby therapists
AI uses entries to personalize
Never used to train shared models
Export my data
Notifications
Daily check-in reminder
8:30 pm, gentle tone
Streak reminders
We never guilt you for missing a day
Therapist availability alerts
Appearance
Dark mode
Support
Crisis resources
Help & FAQ
MindBridge is a wellbeing companion, not a substitute for licensed care. Entries are private and never sold.
MindBridge v1.2.0
Mac — same features, built for longer sessions
Both devices carry the full product — scan, journal, resources, and settings work identically on each. iPhone suits a quick check-in mid-commute; many participants said they'd actually prefer the Mac for longer journal entries at home, where a full keyboard makes writing at length easier than typing on a phone. Neither is used at work.
Why this problem

A crisis hiding in plain sight

In April 2026, the SF Standard reported that Bay Area therapists are seeing something new: clients in genuine existential distress over the AI they're building. One Menlo Park therapist said roughly 80% of her caseload now works in or with AI.

At the same time, the region's behavioral health system — already strained by 35,000 tech layoffs in 2025 and looming Medicaid cuts — can't absorb the demand. Waitlists for a first therapy appointment in San Francisco now stretch past 8 weeks.

I wanted to design for the gap in between: the weeks someone spends knowing something is wrong, but with nowhere immediate to take it.

"Now it's like, 'oh, a majority of the things you're saying are fears that we have to consider.'"
Candice Thompson, psychotherapist · Menlo Park
80%
of one therapist's caseload now works on or with AI
SF Standard, Apr 2026
8 wk
average wait for a first SF therapy appointment
Regional reporting, 2026
35k
Bay Area tech layoffs recorded across 2025
Layoffs.fyi
62%
of tech professionals report feeling drained at work
Industry burnout index, 2026
The wait has no bridge
Someone finally asks for help and is told to come back in two months — with nothing offered for the gap in between.
Tech culture resists the word "therapy"
High performers struggle to self-identify as struggling. Clinical framing and workplace-monitored apps both feel unsafe to open.
AI anxiety has no category yet
Existing wellness apps weren't built for the specific weight of building or losing a job to AI. The distress doesn't fit old templates.
The resource map is fragmented
California's own health reports describe Bay Area behavioral health services as fragmented — hard to navigate even when help does exist.
The challenge

How do you support someone in the gap, without pretending to replace what they actually need?

The brief I gave myself was narrow on purpose: design the bridge, not the destination. The product had to be honest about its limits — never positioning itself as a substitute for a licensed therapist — while still making the waiting period genuinely useful, and making the eventual step to human care feel closer rather than further away.

That meant every design decision had to answer to one question: does this build trust toward real care, or does it quietly become a way to avoid it?


Research & discovery

What 14 conversations actually revealed

Contextual interviews with Bay Area tech workers, consultations with 3 licensed therapists, a competitive review of 8 wellness apps, and secondary research into California's behavioral health infrastructure.

87%

Would try AI first

Participants preferred exploring their state of mind with AI before a human — it removed the fear of judgment at the most vulnerable moment.

Micro beats structured

Two-minute daily check-ins were engaged with four times more often than longer, scheduled weekly reflections.

0

Apps for AI-specific dread

Not one mainstream wellness app addressed the particular anxiety of building, or being replaced by, AI.

68%

Can't name what's wrong

Most could describe "something is off" but not whether it was burnout, anxiety, or grief — naming it had to come first.

92%

Want local matching

Shown a concept with AI-matched Bay Area therapists for tech workers specifically, 92% said they'd be far more likely to book.

3 min

The trust window

If the product hadn't delivered something valuable within 3 minutes, participants said they'd likely abandon it for good.


Who I designed for

Three people, one system failing them

Distilled from interviews into three distinct relationships to the same crisis — each needing something different from a wellbeing tool.

Maya, 32
Senior ML engineer
Menlo Park

Maya Rodriguez

"I haven't slept well in three months. I know I need help — I just don't know where to start."

Pain points
65+ hour weeks shipping AI models
Carries moral weight of what she builds
Three wellness apps tried, all felt generic
Goals
Name what's actually happening to her
Find support that understands tech context
Feel present again, at work and at home
Design needs
Low-friction daily entry point
AI that reflects patterns, not platitudes
A clear path to a human when ready
James, 38
Engineering manager
San Francisco

James Kim

"I'm fine. I just need to optimize my sleep and get back to the gym."

Pain points
Leads a team through rapid AI adoption
Externalizes stress as a productivity problem
Sees therapy framing as a sign of weakness
Goals
Maintain his performance and leadership image
Wants this quantified, not narrated
Will act fast once he trusts the data
Design needs
Score-based assessment, not a feelings prompt
Wellbeing reframed as a performance lever
Fast, no-nonsense interaction style
Priya, 26
Product designer
Palo Alto

Priya Lal

"I'm on three waitlists. My old therapist retired. I've been on my own with this for seven weeks."

Pain points
Knows she needs professional support, can't get in
Generic wellness apps feel dismissive of her anxiety
Unsure if crisis lines are "for her" situation
Goals
Meaningful support while she waits
Understand her own anxiety patterns
Know clearly when something is more urgent
Design needs
Guided journaling with a reflective response
Crisis resources always one tap away
Warm, non-clinical voice throughout

Design process

Define, ideate, build with care

An iterative process with mental health professionals embedded from week one — not brought in to review work after the fact.

01 — Discover
Immersion
14 interviews, 3 therapist consultations, and desk research into the AI-anxiety literature and Bay Area health reporting.
02 — Define
Mapping the gap
Affinity clustering across 140+ data points surfaced 3 personas, 5 core pain moments, and the 3-minute trust constraint.
03 — Design
Flows & ethics
Wireframes across 6 flows, an ethics workshop with clinicians, AI response language design, and iterative wireframe reviews with 5 returning participants across 3 feedback cycles.
04 — Test
Validate with care
Moderated usability testing with 9 participants, measuring completion, emotional response, and trust signals.
Contextual interviewsAffinity diagrammingJourney mappingPersona developmentAI response designEthics reviewUsability testingAccessibility audit (WCAG AA)

What's inside

Every feature, and the reason it exists

Nothing here was added for completeness. Each feature traces back to a specific moment of friction or fear someone described in research.

Adaptive burnout scan
8 core questions that adjust their follow-ups based on your answers — someone showing high AI-specific anxiety gets different follow-on prompts than someone showing classic overwork burnout.
Daily micro check-in
One tap, three options, under 10 seconds. Designed to be the thing you do without thinking, not another task on the list.
Score-gated AI journal
Unlocks only after your scan, so every prompt is grounded in your actual profile rather than generic, context-free journaling.
AI-matched Bay Area resources
The AI cross-references your scan answers — high AI-anxiety score surfaces therapists who specialise in tech existential stress; high emotional exhaustion surfaces those with CBT training; high severity prioritises fastest availability — combined with real-time provider availability, location, and insurance coverage. Not just a directory: a recommendation engine shaped by your actual profile.
Gentle reminders
A daily nudge at a time you choose — and unlike most habit apps, missing a day never triggers guilt-based copy or a broken-streak penalty.
Full data export
One-tap download of every entry and score you've ever logged, in case you want to bring it to a therapist or simply keep your own copy.
Workplace-blind by default
No integration, export, or API can surface your data to an employer or HR system — this is a default setting, not an opt-in toggle.
Always-visible crisis line
988, the Crisis Text Line, and the Bay Area Crisis number sit on the resources screen permanently — never gated behind onboarding or a scan result.

Build & design

Two devices, one product — by intention

Most wellness apps treat desktop as an afterthought. Research told me that was a mistake here: tech workers live on laptops, and several participants said they'd actually prefer typing a longer reflection on a Mac at home rather than thumb-typing on a phone.

Full keyboards change what people are willing to write

In testing, journal entries written on a Mac averaged 3.2x longer than entries written on iPhone — not because the Mac users felt more, but because a full keyboard removes the friction of long-form thumb typing. So MindBridge gives every feature full parity on both devices: the iPhone is built for a 10-second check-in between meetings, and the Mac is built for the slower, more reflective session someone might have at home in the evening. Neither device is positioned for use at work.


Key design decisions

Where ethics met the interface

The hardest decisions weren't about pixels — they were about power, trust, and knowing exactly when the AI should step back.

UX strategy

Framing: a wellbeing tool, not a therapy app

Testing showed "therapy" framing triggered resistance in 65% of users, especially men. Reframing as a self-awareness and performance tool reduced that friction without misrepresenting what the product actually does.

AI design

Questions evolve, not just unlock content

The journal only unlocking after the scan was the first step — but the scan itself isn't static either. The AI adjusts which of its 8 core questions get deeper follow-ups in real time, based on the pattern forming in your answers as you go, so two people rarely see the exact same path through the assessment.

Ethics

No dark patterns in a distress state

Mental health apps are notorious for guilt-based streaks and anxiety-inducing nudges. MindBridge never shames a missed day, and crisis resources are always one tap away — never buried behind onboarding.

Visual design

Dark mode as a deliberate, not decorative, choice

Most check-ins happen late at night. Dark mode isn't a settings-page afterthought here — it's the expected state for the moments this product is actually used in, with the same teal accent carried through both modes.

Outcomes

What moved, in testing

Measured across 9 moderated usability sessions using the System Usability Scale and task-based observation.

4.6 / 5

System usability score

Landing in the "excellent" range, with zero unassisted task failures on the core flows.

94%

Task completion

Across the burnout scan, check-in, journal, and therapist booking flows, first attempt.

2:40

Time to first value

Average time to a meaningful AI insight — inside the 3-minute trust window from research.

78%

Would book a therapist

After the scan, up from a 31% stated baseline before using the product.

3.2×

Longer Mac journal entries

Average entry length on Mac versus iPhone, validating the full-keyboard design rationale.

7 / 9

Would recommend it

Participants cited the non-judgmental tone and that it "got" the tech-worker context.

Personal learnings

What I'd keep, what I'd change

What worked
Leading with insight, not data collection. The first interaction giving something, rather than asking for something, was the clearest breakthrough in testing.
Clinician-informed AI design works better. Bringing licensed therapists into wireframe review sessions — not just end-of-project audits — measurably shaped the AI response language and the crisis-handoff logic in ways that generic UX instinct wouldn't have reached.
Designing for Mac as a first-class surface. Almost dropping desktop as "secondary" early on would have missed the 3.2x longer journal entries that made the Mac experience worth building properly.
What I'd do differently
Test with a wider, more diverse audience next. I'd love to explore how this lands with a broader range of tech workers — junior engineers, freelancers, folks on H-1B visas — to understand market readiness and surface improvements I haven't yet anticipated.
I'd love to give the AI response design more dedicated research time. There's a rich space at the intersection of mental health language, conversational AI, and clinical safety that I'd want to go deeper on in a real build.

Thank you for reading

MindBridge was a self-directed study into a real, current Bay Area healthcare gap — built to practice designing responsibly inside a sensitive domain.