Build a demo app called "Candid.ai" — an AI-powered qualitative research interview platform that conducts real consumer interviews at scale. Companies define their target customer profile and research objective, Candid recruits matched participants from a vetted panel of 100,000+ consumers, then an AI interviewer conducts natural one-on-one voice conversations — probing deeper on interesting answers, skipping irrelevant questions, reading emotional cues, and capturing nuance that surveys never could. Findings are automatically synthesised into structured insight reports with sentiment analysis, theme clustering, and direct quote evidence — delivered in 48 hours, not 6 weeks. It costs a fraction of traditional qual research.
WHO: Product managers, UX researchers, brand strategists, and insight leads at consumer brands who need qualitative customer understanding but can't afford the time or cost of traditional research.
WHAT: An AI qualitative research platform that recruits participants, conducts adaptive voice interviews, and synthesises findings into boardroom-ready reports.
WHY: Traditional qual research takes 6-8 weeks and costs $30-80K per study. Candid delivers the same depth in 48 hours for under $5K. Most companies skip qual entirely because of cost — Candid makes it accessible to every product decision.
This is a DEMO app. Use hardcoded dummy data throughout. No real authentication — clicking "Try the Demo" or "Explore Dashboard" on the landing page navigates directly into the app dashboard. No login screen, no signup forms that actually validate. The simulated brand inside the dashboard is a direct-to-consumer meal kit delivery company called "Kindred" — they use Candid.ai to run qualitative research studies on their customers.
VISUAL STYLE:
Apply a dark mode SaaS dashboard aesthetic throughout — inspired by Linear, Notion, and Dovetail. Background: near-black (#09090F). Card surfaces: dark charcoal (#13131A) with subtle 1px borders (#1C1C28). Primary accent color: indigo/violet (#7C5CFC). Secondary accent: teal (#14B8A6) for positive sentiment and success states. Tertiary accent: soft amber (#F59E0B) for mixed/neutral sentiment. Quaternary accent: rose (#F43F5E) for negative sentiment. Text: white (#F5F5F7) for headings, light gray (#A1A1AA) for secondary text. Use the Inter font family throughout. Cards should use subtle glassmorphism — very slight background blur (backdrop-filter: blur(12px)) and transparency (rgba(19,19,26,0.8)). Rounded corners (14px) on all cards and containers. Generous spacing — 24px padding inside cards, 20px gap between grid items. Status badges use semantic colors: teal (#14B8A6) for completed/positive, amber (#F59E0B) for in-progress/neutral, rose (#F43F5E) for issues/negative, indigo (#7C5CFC) for active/primary, blue (#3B82F6) for info. Use subtle indigo-to-purple gradient accents (#7C5CFC to #A78BFA) on hero sections, key CTAs, and primary action buttons. Hover effects on cards should slightly brighten borders. Buttons have a slight scale(1.02) on hover.
LOGO: The Candid.ai logo should be the text "candid" in lowercase, bold Inter font, white — followed by ".ai" in the indigo accent color (#7C5CFC). Use this in the nav, sidebar, and footer.
PAGES AND ROUTES:
| Route |
Page |
| / |
Landing Page (marketing) |
| /app/dashboard |
Dashboard |
| /app/projects |
Research Projects |
| /app/study |
Study Detail View |
| /app/participants |
Participant Panel |
| /app/live |
Live Interviews |
The landing page (/) has its own layout — no sidebar, just a transparent nav bar.
All /app/* pages share an app layout with a fixed left sidebar (hidden on mobile, becomes a slide-in overlay).
PAGE 1: LANDING PAGE (route: "/")
Fixed Navigation Bar:
- Transparent background with backdrop blur, border-bottom rgba(255,255,255,0.05)
- Logo on the left
- Nav links center: "Features", "How It Works", "Testimonials" (smooth scroll to sections)
- "Try the Demo" button on the right (links to /app/dashboard) — indigo gradient background (#7C5CFC to #A78BFA), white text
- Mobile: hamburger menu that opens a dropdown with the same links
Hero Section:
- Full viewport height, centered content
- Background: near-black with a subtle radial gradient glow of indigo/violet at ~8% opacity behind the headline area
- Eyebrow text above headline: "AI QUALITATIVE RESEARCH" in small caps, indigo color, letter-spacing 3px
- Large headline: "Talk to 100 customers in the time it takes to schedule one." — text-5xl on mobile, text-7xl on desktop, bold, tight tracking
- Subheadline in muted gray (max-w-2xl): "Candid.ai recruits real participants, conducts adaptive voice interviews, and delivers boardroom-ready insight reports in 48 hours — for a fraction of the cost of traditional qualitative research."
- Two CTA buttons side by side:
- "Try the Demo" (primary, indigo gradient background, white text) — navigates to /app/dashboard
- "See How It Works" (secondary, transparent with white/10 border) — scrolls down to How It Works section
- Below buttons, small muted text: "Used by 200+ product teams · 50,000+ interviews completed · 48-hour average turnaround"
- Floating stats card below: glass card with 4 stats in a row: "100K+" (indigo, "Vetted Panellists"), "48hrs" (teal, "Average Turnaround"), "92%" (amber, "Insight Accuracy Rating"), "1/10th" (rose, "The Cost of Traditional")
- Use framer-motion staggered fade-in animations (initial opacity:0, y:20)
How It Works Section (id="how-it-works"):
- Section heading: "From research brief to boardroom report in 48 hours" + subtitle "Three steps. No recruitment headaches. No weeks of waiting."
- 3 glass cards in a row, each with:
- Step number label ("Step 1", "Step 2", "Step 3")
- Colored icon in a rounded square (icon background at 10% opacity of the icon color)
- Step 1: Target icon (indigo) — "Define & Recruit" — "Tell us who you need to talk to — demographics, behaviours, attitudes. Our AI matches and recruits from a vetted panel of 100,000+ real consumers. Participants are screened, scheduled, and ready within hours."
- Step 2: Mic icon (teal) — "AI Interviews" — "Our AI conducts natural, adaptive one-on-one voice conversations. It probes deeper when something interesting surfaces, skips irrelevant questions, reads emotional cues, and captures the nuance that a checkbox survey never could."
- Step 3: FileText icon (amber) — "Synthesise & Report" — "Interviews are automatically transcribed, analysed, and synthesised into structured insight reports — complete with sentiment analysis, theme clustering, and direct quote evidence. Delivered to your inbox in 48 hours."
- Animate in on scroll with stagger
Features Section (id="features"):
- Heading: "Qualitative depth at quantitative scale" + subtitle "Everything you need to understand your customers — without the 6-week wait."
- 2x2 grid of feature cards (glass cards):
- Card 1: Users icon (indigo) — "Smart Participant Recruitment" — "Define your ideal participant profile and our AI matches from a vetted panel of 100,000+ consumers. Demographic filters, behavioural screening, and attitude matching — participants are recruited and scheduled automatically."
- Card 2: Mic icon (teal) — "Adaptive AI Interviews" — "Our AI interviewer conducts natural voice conversations that feel human. It follows up on interesting answers, probes for underlying motivations, skips irrelevant paths, and adapts its approach based on each participant's responses."
- Card 3: Brain icon (amber) — "Automated Synthesis" — "Raw interviews are transcribed, coded, and synthesised into structured reports automatically. Theme clustering, sentiment analysis, contradiction detection, and emergent insight identification — all without a human analyst."
- Card 4: BarChart3 icon (rose) — "Boardroom-Ready Reports" — "Delivered in 48 hours: executive summaries, key themes with supporting quotes, sentiment breakdowns, persona clusters, and actionable recommendations. Export to PDF, share a live link, or present directly."
Testimonials Section (id="testimonials"):
- Heading: "Teams that switched to Candid never go back" + subtitle
- 3 glass testimonial cards:
- Quote icon at top (indigo at 30% opacity)
- Testimonial 1: "We used to budget $60K and 8 weeks for a single qual study. Candid gave us deeper insights in 2 days for $3K. We've now run 12 studies this quarter — we'd have done one the old way." — Maya Ishikawa, VP of Product, Relay (fintech)
- Testimonial 2: "The AI interviewer picked up on something our human moderators missed for years — a specific moment in onboarding where users feel anxious, not confused. That distinction changed our entire redesign." — Liam Okonkwo, Head of UX Research, Bloom Health
- Testimonial 3: "I was sceptical about AI-conducted interviews. Then I read the transcripts. The follow-up questions were better than most junior researchers I've hired. The probing on emotional responses was genuinely impressive." — Claire Dubois, Chief Insights Officer, Terrain Outdoors
CTA Section:
- Glass card with gradient border, large padding
- Background glow effect inside
- "See Candid.ai in action"
- Subtext: "Explore the demo dashboard for Kindred, a meal kit delivery brand running qualitative research studies on their customers."
- "Explore Demo Dashboard" button (indigo gradient) — navigates to /app/dashboard
Footer:
- Border-top, logo on left
- Links center: Features, How It Works, Pricing, Privacy, Terms (all # hrefs)
- Tagline: "AI-powered qualitative research. Deeper insights. Fraction of the time."
- Copyright: "2026 Candid.ai. All rights reserved."
APP LAYOUT (shared by all /app/* pages):
Sidebar (Desktop):
- Fixed, left side, 240px wide, full height
- Background: #09090F, right border rgba(255,255,255,0.05)
- Logo at top
- Navigation items with Lucide icons, rounded-xl hover/active states:
- Dashboard (LayoutDashboard icon) → /app/dashboard
- Projects (FolderKanban icon) → /app/projects
- Study Detail (FileSearch icon) → /app/study
- Participants (Users icon) → /app/participants
- Live Interviews (Radio icon) → /app/live
- Active item: indigo left border (3px), bg-white/10, text-white
- Inactive: text-muted-foreground hover:bg-white/5
- Divider line above bottom section
- Organisation info at bottom: indigo circle with "K" initial, "Kindred" company name, "Growth Plan" badge in small indigo text
Sidebar (Mobile):
- Hidden by default
- Mobile header bar: fixed top, hamburger button + logo
- Hamburger opens a slide-in overlay (dark backdrop + sidebar from left)
- Close button in the overlay
Main Content Area:
- Left padding for sidebar on desktop (pl-64), top padding for mobile header on mobile (pt-14)
- Content max-width: max-w-6xl, centered, with padding
PAGE 2: DASHBOARD (route: "/app/dashboard")
Header:
- "Welcome back, Kindred" (h1, bold, white)
- "Here's your research activity for February 2026." (muted text)
4 Key Metric Cards (responsive grid: 1→2→4 columns):
Each card is a glass card with:
- Label (muted, small)
- Icon in a colored rounded square (top right)
- Large number
- Small change indicator
Cards:
- "Active Studies" — "4" — FolderKanban icon (indigo bg) — "2 completed, 1 in field, 1 recruiting"
- "Interviews Completed" — "312" — Mic icon (teal bg) — "↑ 84 this week"
- "Avg. Turnaround" — "41 hrs" — Clock icon (amber bg) — "↓ 7 hrs vs last month"
- "Participants Recruited" — "347" — Users icon (blue bg) — "98.3% show-up rate"
Performance Snapshot Card (full width glass card):
- Title: "This Month at a Glance"
- A row of mini-stats with small icons and coloured text:
- Studies launched: 3
- Interviews conducted: 156
- Avg. interview duration: 24 min
- Themes identified: 47
- Reports delivered: 2
- Client satisfaction: 4.8/5.0
Main Content Grid (3/5 + 2/5 on desktop):
Left: Recent Activity Feed (glass card):
- Header: "Recent Activity" with "View All" link
- 7 activity items, each with coloured icon, description text, and relative timestamp:
- FileText (teal) — "Report delivered for 'Why Customers Cancel After Month 3' — 47 key findings synthesised from 80 interviews" — 2 hours ago
- Mic (indigo) — "Interview completed — Hannah Byrne (32, Dublin) discussed meal variety and recipe fatigue for the 'Churn Drivers' study" — 3 hours ago
- Users (blue) — "12 new participants recruited for 'Premium Tier Willingness to Pay' — matching criteria: household income >$120K, current subscribers" — 5 hours ago
- Mic (indigo) — "Interview completed — Marcus Obi (28, London) gave strong feedback on packaging sustainability for the 'Eco Packaging' study" — 6 hours ago
- AlertCircle (amber) — "Emerging theme detected in 'Churn Drivers' study — 'recipe repetition' mentioned by 73% of participants unprompted" — 8 hours ago
- CheckCircle (teal) — "Study 'Competitor Switching Motivations' marked as complete — 80/80 interviews finished, report generation started" — 12 hours ago
- Radio (rose) — "Live interview in progress — James Whitfield (45, Manchester) discussing family meal planning for 'New Menu Categories' study" — Now
Right: Quick Actions (glass card):
- Header: "Quick Actions"
- 3 action cards stacked:
- FolderKanban icon (indigo) — "View Projects" — "4 active studies" — ArrowRight, links to /app/projects
- FileSearch icon (teal) — "Latest Report" — "Churn Drivers — ready" — ArrowRight, links to /app/study
- Radio icon (rose) — "Live Interviews" — "3 in progress now" — ArrowRight, links to /app/live
PAGE 3: RESEARCH PROJECTS (route: "/app/projects")
Header:
- "Research Projects" (h1) + "All qualitative research studies for Kindred." (muted)
Filter bar: 2 dropdowns:
- Status: All / Recruiting / In Field / Analysis / Completed
- Research Type: All / Exploratory / Evaluative / Strategic
All filtering is client-side using state.
Project Cards (list of glass cards):
Project 1:
- Title: "Why Customers Cancel After Month 3"
- Status: Teal badge "COMPLETED"
- Research type: "Exploratory" (muted label)
- Objective: "Understand the specific triggers and experiences that cause subscribers to cancel between months 3-6, with focus on whether the issue is product, price, or habit."
- Target participants: "Cancelled subscribers, 25-45, subscribed for 3-6 months before cancelling"
- Stats row: Interviews: 80/80 | Avg. Duration: 26 min | Themes Found: 18 | Sentiment: Mixed (amber dot)
- Turnaround: "Completed in 44 hours"
- "View Report" button — links to /app/study
Project 2:
- Title: "Competitor Switching Motivations"
- Status: Teal badge "COMPLETED"
- Research type: "Strategic"
- Objective: "Identify why customers switch from Kindred to HelloFresh, Gousto, or other competitors — and what would win them back."
- Target participants: "Former Kindred subscribers who switched to a competitor, 28-50, cancelled within last 6 months"
- Stats row: Interviews: 80/80 | Avg. Duration: 22 min | Themes Found: 14 | Sentiment: Negative (rose dot)
- Turnaround: "Completed in 38 hours"
- "View Report" button
Project 3:
- Title: "Premium Tier Willingness to Pay"
- Status: Indigo badge "IN FIELD"
- Research type: "Evaluative"
- Objective: "Test appetite for a premium subscription tier ($79/week) featuring organic ingredients, chef-designed recipes, and wine pairings. Understand perceived value and price sensitivity."
- Target participants: "Current subscribers 6+ months, household income >$120K, interested in premium food"
- Stats row: Interviews: 48/75 | Avg. Duration: 28 min | Themes Found: 11 (emerging) | Sentiment: Positive (teal dot)
- Progress bar: 64% complete
- "View Progress" button
Project 4:
- Title: "Eco Packaging Perception Study"
- Status: Amber badge "RECRUITING"
- Research type: "Evaluative"
- Objective: "Gauge customer reaction to proposed switch from plastic insulation to wool-based packaging — willingness to accept trade-offs on delivery window and shelf life."
- Target participants: "Active subscribers 3+ months, mix of eco-conscious and convenience-driven, 25-55"
- Stats row: Recruited: 32/60 | Screening: 18 pending | Start date: 28 Feb 2026
- "View Recruitment" button
PAGE 4: STUDY DETAIL VIEW (route: "/app/study")
This page shows the completed study "Why Customers Cancel After Month 3" in detail. This is the most important page — it demonstrates the core value of the platform.
Header:
- Breadcrumb: "Projects / Why Customers Cancel After Month 3"
- Title: "Why Customers Cancel After Month 3" (h1)
- Status: Teal badge "COMPLETED" + "Report delivered 2 hours ago"
- Study metadata row: "80 interviews · 26 min avg · 18 themes · 44-hour turnaround"
Tab bar: "Executive Summary" (default active) | "Key Themes" | "Quotes & Evidence" | "Sentiment Analysis"
EXECUTIVE SUMMARY TAB:
Executive Summary Card (full width glass card):
- "Executive Summary" heading with a small Brain icon
- Paragraph text: "Across 80 interviews with cancelled subscribers (months 3-6), three dominant cancellation drivers emerged: recipe repetition and declining novelty (cited by 73% of participants), portion size dissatisfaction (cited by 58%), and a perceived drop in ingredient quality after an initial 'honeymoon' period (cited by 41%). Notably, price was rarely the primary driver — only 12% cited cost as their main reason for leaving. The strongest predictor of cancellation was the moment a subscriber cooked a recipe they had already received, which 67% described as 'the beginning of the end.' Participants consistently expressed that they joined Kindred for inspiration and variety, and cancellation was triggered when that promise felt broken."
Key Findings Cards (3 glass cards in a row):
- Finding 1: Sparkle icon (rose) — "Recipe Fatigue is the #1 Killer" — "73% of cancellers cited recipe repetition as their primary frustration. Most reported receiving a duplicate recipe within 8-10 weeks. The emotional response was disproportionately negative — participants used words like 'betrayed,' 'lazy,' and 'insulting.'"
- Finding 2: Sparkle icon (amber) — "Portions Don't Scale for Families" — "58% said portions were too small for families with children over 10. The 2-person and 4-person plans were seen as '2 adults' and '2 adults + 2 small children.' Several participants described supplementing Kindred meals with additional cooking."
- Finding 3: Sparkle icon (indigo) — "Quality Perception Drops at Week 8" — "41% perceived a decline in ingredient quality after the first 2 months. Whether real or psychological, the 'honeymoon' period creates an expectation that later boxes don't meet. Participants specifically mentioned 'thinner cuts of meat' and 'less fresh herbs.'"
Actionable Recommendations Card (glass card):
- Heading: "AI-Generated Recommendations"
- 5 recommendation items with numbered circles:
- "Expand the recipe library by 40%+ and guarantee no repeats within the first 6 months. This single change addresses the top cancellation driver."
- "Introduce a 'Family+' plan with genuinely larger portions designed for households with teenagers, not just scaling up the 2-person portions."
- "Implement a 'Week 6 check-in' — a proactive customer touchpoint before the quality perception drop that acknowledges their experience and offers customisation."
- "Add a 'surprise me' recipe option each week — something outside the subscriber's usual cuisine preferences — to combat novelty fatigue."
- "Create a transparent ingredient sourcing page so subscribers can verify quality standards aren't changing — addressing the perception gap directly."
KEY THEMES TAB:
Theme cards (list of glass cards, each with a colour-coded sentiment bar on the left):
Theme 1:
- Rose sentiment bar (left edge)
- Theme name: "Recipe Repetition & Novelty Fatigue" — Mentioned by: 73% (58/80 participants)
- Sentiment: rose badge "Strongly Negative"
- Summary: "Participants expected an ever-expanding menu and were emotionally disappointed when they received a recipe they'd already cooked. This was consistently described as the moment they started considering cancellation."
- Supporting quotes (2-3 short quotes in muted italic):
- "I got the Thai green curry again in week 9. I literally said to my partner, 'I think we're done with Kindred.'" — Sophie A., 34, Bristol
- "The whole point was not having to think about what to cook. Once I'd seen it all before, I was just paying someone to repeat my own cooking." — David L., 41, Edinburgh
Theme 2:
- Amber sentiment bar
- Theme name: "Portion Size Mismatch" — Mentioned by: 58% (46/80 participants)
- Sentiment: amber badge "Mixed / Frustrated"
- Summary: "The 4-person plan was perceived as adequate for 2 adults and 2 young children, but insufficient for families with teenagers or larger appetites. Participants felt misled by the plan description."
- Supporting quotes:
- "My 14-year-old eats more than I do. The 'family' plan is really a '2 adults and 2 toddlers' plan." — James T., 44, Manchester
- "We'd finish the meal and then I'd have to make toast or something. That's not a meal kit — that's a starter kit." — Priya S., 38, Birmingham
Theme 3:
- Amber sentiment bar
- Theme name: "Ingredient Quality Perception Decline" — Mentioned by: 41% (33/80 participants)
- Sentiment: amber badge "Disappointed"
- Summary: "Whether objectively true or not, participants perceived a decline in ingredient quality after the first 2 months. The initial boxes set a high standard that subsequent deliveries seemed to fall short of."
- Supporting quotes:
- "The first month, I was getting beautiful thick-cut salmon fillets. By month 4 they were paper-thin. Maybe it's seasonal, but it felt like a downgrade." — Rachel H., 29, London
- "The herbs went from being these generous bunches to sad little plastic packets. I noticed." — Tom O., 36, Leeds
Theme 4:
- Teal sentiment bar
- Theme name: "Price Not the Primary Issue" — Mentioned by: 12% (10/80 participants)
- Sentiment: teal badge "Neutral / Accepting"
- Summary: "Only a small minority cited price as their primary cancellation reason. Most acknowledged the price was fair for the value when the product was good — the issue was that the product experience declined."
- Supporting quotes:
- "Honestly, the price was fine. I'd happily pay even more if the recipes stayed interesting. I was paying for inspiration, not ingredients." — Olivia M., 31, Glasgow
- "If they'd kept it fresh, I'd still be subscribed at double the price." — Noah K., 27, Cardiff
Theme 5:
- Indigo sentiment bar
- Theme name: "The 'Beginning of the End' Moment" — Mentioned by: 67% (54/80 participants)
- Sentiment: indigo badge "Emotionally Significant"
- Summary: "A clear pattern emerged: the moment a subscriber receives a recipe they've already had before is a psychological turning point. It shatters the illusion of infinite variety and triggers active evaluation of whether to continue."
- Supporting quotes:
- "It's funny — I could point to the exact meal. It was the second time I got the chicken shawarma bowl. I remember thinking, 'wait, have I already made this?' And then I started thinking about whether this was actually worth it." — Grace W., 33, Brighton
- "One repeat wouldn't have killed me. But it made me look back and realise I'd only had maybe 30 unique meals in 4 months. That maths didn't work for me." — Daniel F., 39, Newcastle
QUOTES & EVIDENCE TAB:
A searchable list of direct participant quotes, each as a glass card:
Quote 1:
- Avatar "SA" with teal border — Sophie A., 34, Bristol — Cancelled after 4 months
- Theme tag: "Recipe Repetition" (rose badge)
- Quote: "I got the Thai green curry again in week 9. I literally said to my partner, 'I think we're done with Kindred.' It's not that one repeated recipe is the end of the world — it's that it broke the spell. Suddenly I was thinking about all the things I didn't like instead of all the things I did."
- Interviewer follow-up shown: AI asked: "You said it 'broke the spell.' Can you tell me more about what that spell was?" — Response: "Yeah, like... Kindred made me feel like this adventurous home cook. Like I was discovering new cuisines every week. The repeat made me feel like I was just reheating someone else's ideas."
- Sentiment: Rose dot — "Negative — Disappointment/Betrayal"
Quote 2:
- Avatar "JT" with amber border — James T., 44, Manchester — Cancelled after 5 months
- Theme tag: "Portion Size" (amber badge)
- Quote: "My 14-year-old eats more than I do. The 'family' plan is really a '2 adults and 2 toddlers' plan. I actually weighed the chicken once — 400 grams for 4 people? That's 100 grams each. My son would eat that as a snack."
- Sentiment: Amber dot — "Negative — Frustration"
Quote 3:
- Avatar "RH" with amber border — Rachel H., 29, London — Cancelled after 3 months
- Theme tag: "Quality Decline" (amber badge)
- Quote: "The first month, I was getting beautiful thick-cut salmon fillets. By month 4 they were paper-thin. Maybe it's seasonal, but it felt like a downgrade. And once you start noticing one thing, you notice everything — the veg wasn't as crisp, the packaging felt cheaper."
- Sentiment: Amber dot — "Negative — Disappointed"
Quote 4:
- Avatar "OM" with teal border — Olivia M., 31, Glasgow — Cancelled after 6 months
- Theme tag: "Price Acceptance" (teal badge)
- Quote: "Honestly, the price was fine. I'd happily pay even more if the recipes stayed interesting. I was paying for inspiration, not ingredients. The moment it stopped being inspiring, the price became visible — if that makes sense."
- Interviewer follow-up: AI asked: "What would 'staying inspiring' look like for you?" — Response: "More cuisines I've never tried. I want to be slightly out of my depth. I want to text my friends about what I made, not scroll past it."
- Sentiment: Teal dot — "Neutral — Constructive"
Quote 5:
- Avatar "GW" with indigo border — Grace W., 33, Brighton — Cancelled after 4 months
- Theme tag: "Turning Point" (indigo badge)
- Quote: "It's funny — I could point to the exact meal. It was the second time I got the chicken shawarma bowl. I remember thinking, 'wait, have I already made this?' And then I started thinking about whether this was actually worth it. That single meal turned me from a happy subscriber into someone who was actively evaluating."
- Sentiment: Indigo dot — "Significant — Emotional Turning Point"
Quote 6:
- Avatar "DL" with rose border — David L., 41, Edinburgh — Cancelled after 3 months
- Theme tag: "Recipe Repetition" (rose badge)
- Quote: "The whole point was not having to think about what to cook. Once I'd seen it all before, I was just paying someone to repeat my own cooking. I can do that for free."
- Sentiment: Rose dot — "Negative — Dismissive"
SENTIMENT ANALYSIS TAB:
Overall Sentiment Card (glass card):
- Donut chart showing: Negative 52%, Mixed/Neutral 31%, Positive 17%
- Center text: "80 interviews analysed"
- Label: "Overall study sentiment skews negative, as expected for a churn analysis study."
Sentiment by Theme Card (glass card, horizontal bars):
- Recipe Repetition: 89% negative, 8% neutral, 3% positive (rose bar dominant)
- Portion Size: 71% negative, 22% neutral, 7% positive (rose bar dominant)
- Quality Decline: 64% negative, 28% neutral, 8% positive (amber/rose mix)
- Price Perception: 18% negative, 34% neutral, 48% positive (teal bar dominant)
- Turning Point Moment: 78% negative, 19% neutral, 3% positive (rose bar dominant)
Emotional Intensity Card (glass card):
- "Words participants used most frequently (weighted by emotional intensity):"
- Word cloud or tag list: "repetitive" (large, rose), "disappointing" (large, rose), "lazy" (medium, rose), "betrayed" (medium, rose), "small portions" (large, amber), "not enough" (medium, amber), "loved at first" (medium, teal), "would come back" (small, teal), "miss it" (small, teal), "insulting" (medium, rose), "downgrade" (medium, amber)
Sentiment Over Interview Timeline Card (glass card):
- Description: "Average sentiment shifts during the course of each interview"
- A simple area chart or timeline visual showing: Opening (neutral) → Early Memories (positive — "participants often started by recalling what they loved about Kindred") → Trigger Discussion (sharp negative — "the moment they discussed why they cancelled") → Solutions Discussion (cautiously positive — "most expressed willingness to return if issues were fixed")
PAGE 5: PARTICIPANT PANEL (route: "/app/participants")
Header:
- "Participant Panel" (h1) + "Vetted research participants matched to your studies." (muted)
Summary Stats Row (4 small glass cards):
- "Total Panel Access" — "100,000+" — Indigo accent
- "Matched for Kindred" — "2,847" — Teal accent
- "Interviewed This Month" — "156" — Blue accent
- "Avg. Show-up Rate" — "98.3%" — Teal accent
Filter bar: 3 dropdowns:
- Study: All / Churn Drivers / Competitor Switching / Premium Tier / Eco Packaging
- Status: All / Completed / Scheduled / Available
- Demographics: All / 25-34 / 35-44 / 45+
All filtering is client-side using state.
Participant Cards (list of glass cards):
Participant 1:
- Avatar "SA" (teal border for completed) + "Sophie A." — Age: 34 — Location: Bristol
- Study: "Why Customers Cancel After Month 3" — Status: teal badge "COMPLETED"
- Profile: "Former Kindred subscriber, 4 months. Marketing coordinator. Lives with partner, no children. Cancelled due to recipe repetition."
- Interview summary: "26 min interview. Highly articulate. Expressed strong emotional attachment to Kindred initially — described cancellation as 'breaking up with a friend.' Key insight: the first repeated recipe was a distinct psychological turning point."
- Key quote: "Kindred made me feel like this adventurous home cook... the repeat made me feel like I was just reheating someone else's ideas."
- Sentiment: Rose dot — "Overall negative, with nostalgic positive undertones"
Participant 2:
- Avatar "JT" (teal border) + "James T." — Age: 44 — Location: Manchester
- Study: "Why Customers Cancel After Month 3" — Status: teal badge "COMPLETED"
- Profile: "Former subscriber, 5 months. IT project manager. Family of 4 (children aged 14 and 11). Cancelled due to portion inadequacy."
- Interview summary: "31 min interview. Practical and specific. Brought detailed examples including weighing portions. Strong advocate for a genuine family plan option."
- Key quote: "The 'family' plan is really a '2 adults and 2 toddlers' plan."
- Sentiment: Amber dot — "Frustrated but rational"
Participant 3:
- Avatar "RH" (teal border) + "Rachel H." — Age: 29 — Location: London
- Study: "Why Customers Cancel After Month 3" — Status: teal badge "COMPLETED"
- Profile: "Former subscriber, 3 months. Graphic designer. Lives alone. Cancelled due to perceived quality decline."
- Interview summary: "22 min interview. Detail-oriented. Noticed specific changes in packaging, ingredient cuts, and herb quantities. Compared early and later boxes with photographic memory."
- Key quote: "Once you start noticing one thing, you notice everything."
- Sentiment: Amber dot — "Disappointed, observant"
Participant 4:
- Avatar "MW" (indigo border for scheduled) + "Maya W." — Age: 37 — Location: Edinburgh
- Study: "Premium Tier Willingness to Pay" — Status: indigo badge "SCHEDULED"
- Profile: "Current subscriber, 9 months. Architect. Household income $145K. Regularly buys organic groceries. Expressed interest in wine pairings in past survey."
- Interview date: "28 Feb 2026, 14:00 GMT"
- Status note: "Pre-interview questionnaire completed. AI interviewer briefed on participant profile."
Participant 5:
- Avatar "KN" (indigo border) + "Kenji N." — Age: 52 — Location: Bath
- Study: "Premium Tier Willingness to Pay" — Status: indigo badge "IN PROGRESS"
- Profile: "Current subscriber, 14 months. Restaurant owner (retired). Household income $180K. Described as 'passionate home cook' in screening."
- Interview duration so far: "18 min (ongoing)"
- Live status: indigo pulse dot — "Interview in progress — discussing willingness to pay for organic certification"
Participant 6:
- Avatar "LB" (amber border for recruiting) + "Lauren B." — Age: 31 — Location: Cardiff
- Study: "Eco Packaging Perception Study" — Status: amber badge "SCREENING"
- Profile: "Active subscriber, 5 months. Environmental scientist. Strong eco-consciousness indicated in screening questionnaire."
- Screening status: "Passed demographic filters. Awaiting attitude screening completion."
PAGE 6: LIVE INTERVIEWS (route: "/app/live")
Header:
- "Live Interviews" (h1) + "AI-conducted interviews happening right now." (muted)
Live Status Bar (full width glass card, indigo gradient border):
- Pulse animation dot (indigo) + "3 interviews in progress" + "12 scheduled today" + "Next interview starts in 22 minutes"
Live Interview Cards (glass cards with subtle indigo glow/border to indicate 'live'):
Interview 1 (Active):
- Indigo pulse dot + "LIVE" badge
- Participant: "Kenji N., 52, Bath" — Study: "Premium Tier Willingness to Pay"
- Duration: "18:42" (incrementing timer display)
- Current topic: "Discussing perceived value of organic ingredients vs. conventional"
- Live transcript excerpt (glass sub-card with monospace font, scrollable, last 5 exchanges):
- AI: "You mentioned you'd be willing to pay more for organic ingredients. Can you walk me through what 'organic' means to you in the context of a meal kit?"
- Kenji: "For me it's about knowing where things come from. I ran a restaurant for 20 years — I know what good produce looks like. Organic is a shortcut for quality when you can't pick it yourself."
- AI: "That's interesting — you're using organic as a proxy for quality rather than for health or environmental reasons. Is that fair?"
- Kenji: "Exactly. I don't care about pesticides honestly. I care about flavour. And organic tomatoes just taste better. That's worth paying for."
- AI: "If Kindred offered a premium tier at $79 per week with guaranteed organic produce, how would you evaluate whether that's worth it compared to your current plan?"
- Sentiment indicator: Teal — "Positive, engaged"
- AI behaviour note: "AI detected strong expertise signal — adapted questions to leverage participant's professional food background"
Interview 2 (Active):
- Indigo pulse dot + "LIVE" badge
- Participant: "Aisha K., 29, London" — Study: "Premium Tier Willingness to Pay"
- Duration: "11:07"
- Current topic: "Exploring wine pairing interest and at-home dining occasions"
- Live transcript excerpt:
- AI: "Tell me about a recent time you cooked a Kindred meal for guests. What was that experience like?"
- Aisha: "Oh, last Saturday actually. I made the lamb tagine for my parents. They were genuinely impressed — my mum asked for the recipe. It felt like I'd actually cooked something special, not just followed instructions."
- AI: "That sounds like a proud moment. How would having a recommended wine pairing included change that experience for you?"
- Aisha: "It would make it feel complete. Like, I know food but I know nothing about wine. Having someone say 'this Moroccan red goes perfectly with this' — that's the kind of thing that makes you feel like an actual host, not someone assembling a box."
- AI: "You used the word 'assembling.' Is that how you sometimes feel about meal kits?"
- Sentiment indicator: Teal — "Positive, emotionally expressive"
- AI behaviour note: "AI probed the word choice 'assembling' — detected potential tension between convenience and authenticity worth exploring"
Interview 3 (Active):
- Amber pulse dot + "LIVE" badge
- Participant: "Ryan M., 36, Dublin" — Study: "Eco Packaging Perception Study" (early pilot)
- Duration: "6:23"
- Current topic: "Initial reactions to wool-based packaging concept"
- Live transcript excerpt:
- AI: "When you hear 'wool-based insulation' instead of the plastic ice packs you currently get — what's your immediate reaction?"
- Ryan: "Honestly? My first thought is whether it'll actually keep things cold. Like, I'm all for sustainability but not if my chicken arrives warm."
- AI: "That's a really practical concern. If I told you the wool insulation keeps contents chilled for up to 48 hours compared to 36 hours for the plastic packs — so actually longer — does that change your reaction?"
- Ryan: "Wait, really? Then yeah, obviously. Why wouldn't they just do that? Unless it costs more and they pass that on."
- AI: "Great question. Let's explore that — if it meant a small price increase, what range would feel acceptable to you?"
- Sentiment indicator: Amber — "Sceptical, warming"
- AI behaviour note: "Participant expressed functional concern before values — AI correctly prioritised addressing practical objection before exploring willingness to pay"
Upcoming Interviews Card (glass card below):
- Heading: "Scheduled Today"
- 4 upcoming items in a compact list:
- "14:00 — Maya W., 37, Edinburgh — Premium Tier — Scheduled" (indigo dot)
- "14:30 — Oliver R., 42, Liverpool — Premium Tier — Scheduled" (indigo dot)
- "15:00 — Freya C., 26, Bristol — Eco Packaging — Scheduled" (indigo dot)
- "15:30 — Nadia P., 33, Birmingham — Eco Packaging — Scheduled" (indigo dot)
Interview Quality Metrics Card (glass card):
- "Today's Interview Quality"
- Avg. depth score: 8.4/10 (indigo bar)
- Avg. follow-up quality: 9.1/10 (teal bar)
- Participant engagement: 92% (teal bar)
- Avg. new insights per interview: 3.2 (indigo bar)
NAVIGATION BEHAVIOR:
- The sidebar should highlight the active page with an indigo left border and slightly brighter background
- Clicking sidebar items navigates between pages
- The landing page is separate from the app pages — it has its own layout without the sidebar
- All "View Details" and navigation links should work and use the router
- The "View Report" button on the first project card in /app/projects should navigate to /app/study
- All buttons labelled as non-functional should look clickable but do nothing or show a subtle tooltip saying "Demo mode"
- Mobile responsive: sidebar collapses to a hamburger menu on small screens
DATA BEHAVIOR:
- All data is hardcoded / static dummy data. No databases, no APIs, no real calculations.
- Numbers, dates, names, and statuses are all fake but realistic.
- The simulated brand is "Kindred" — a direct-to-consumer meal kit delivery company.
- Research studies: "Why Customers Cancel After Month 3" (completed), "Competitor Switching Motivations" (completed), "Premium Tier Willingness to Pay" (in field), "Eco Packaging Perception Study" (recruiting).
- Participant names: Sophie A., James T., Rachel H., Olivia M., Grace W., David L., Tom O., Noah K., Daniel F., Hannah B., Marcus O., Kenji N., Maya W., Aisha K., Ryan M., Lauren B., Priya S., Oliver R., Freya C., Nadia P.
- The app should feel like a real, lived-in product — not an empty shell. Quotes should sound like real people. Insights should be specific and actionable. The data should tell a coherent story across all pages.
IMPORTANT NOTES:
- The landing page should feel premium and conversion-focused — think Intercom or Dovetail marketing pages.
- The dashboard and app pages should feel like a real SaaS product with rich, dense data — think Linear or Notion.
- Charts should use the app colour palette (indigo for primary data, teal for positive, amber for neutral, rose for negative, gray for benchmarks).
- Ensure all data is internally consistent — the same participant names, studies, and metrics should make sense across all pages.
- Make the site fully responsive — mobile, tablet, desktop.
- Use framer-motion sparingly: landing page scroll reveals, animated counters, hover effects on cards. Don't animate page transitions or sidebar toggles.
- All select dropdowns and dialog modals should use shadcn/ui components styled to match the dark theme (dark backgrounds, white/10 borders).
- If any page is too complex, prioritise in this order: (1) Study Detail, (2) Dashboard, (3) Landing Page, (4) Live Interviews, (5) Projects, (6) Participants.
---
## Refinement Prompts
Use these one at a time after the initial app generates. Paste each into Base44's chat.
### Prompt R1 — Design Polish
Refine the visual design across all pages. Make card borders more subtle (reduce opacity). Add a subtle indigo/violet radial glow behind the hero section headline on the landing page. Ensure all status badges are pill-shaped with rounded corners and consistent padding. Add hover effects on all clickable cards that slightly brighten the border color and add a subtle translateY(-2px) lift. Make sure the sidebar has a subtle separator line between the nav items and the organisation profile at the bottom. Ensure the active sidebar item has an indigo left border accent (3px) and slightly brighter background. Sentiment colour coding should be consistent everywhere: teal for positive, amber for mixed/neutral, rose for negative, indigo for significant/noteworthy.
### Prompt R2 — Charts and Visualizations
Add proper chart visualizations. Use Recharts for: On the Study Detail page Sentiment Analysis tab — a donut chart for overall sentiment (52% negative in rose, 31% neutral in amber, 17% positive in teal), horizontal bar charts for sentiment by theme (stacked bars showing negative/neutral/positive proportions). On the Dashboard page, add small sparkline trend indicators inside each of the 4 metric cards showing a positive trend. On the Study Detail Sentiment Analysis tab, add a simple area chart for "Sentiment Over Interview Timeline" showing the emotional arc from neutral → positive → sharp negative → cautiously positive. On the Projects page, add a progress bar on the "in field" study showing 64% completion with an indigo fill.
### Prompt R3 — Landing Page Animations
Add subtle scroll-triggered fade-in animations to the landing page. The hero headline should have a slight fade-up animation on page load (opacity 0 to 1, translateY 20px to 0, over 0.6 seconds). The "How It Works" step cards should stagger in as the user scrolls (each card 150ms after the previous). The feature cards should fade in with a slight scale effect. The floating stats card in the hero should have a count-up number animation when they scroll into view. The testimonial cards should slide in from below. The live interview cards on /app/live should have a subtle indigo pulse glow animation on the border to indicate they are live.
### Prompt R4 — Responsive Layout and Mobile
Ensure full mobile responsiveness. The landing page navigation should collapse to a hamburger menu on screens under 768px. The sidebar on app pages should collapse to a hamburger menu or slide-in overlay on mobile. All card grids (metrics, features, projects) should stack to single column on phone screens. The live transcript excerpts on the Live Interviews page should be horizontally scrollable on mobile. The quote cards on the Study Detail page should stack cleanly with all information visible. Font sizes should scale down gracefully. Touch targets on mobile should be at least 44px. The tab bar on the Study Detail page should be horizontally scrollable on mobile.
### Prompt R5 — Activity Feed and Live Interview Polish
Add small coloured icons next to each activity feed item on the Dashboard — a teal file icon for reports delivered, an indigo mic icon for completed interviews, a blue users icon for recruitment activity, an amber alert icon for emerging themes, and a rose radio icon for live interviews. On the Live Interviews page, add a subtle incrementing timer next to each active interview duration. Add a pulsing indigo dot animation next to the "LIVE" badge on each active interview card. In the Participant Panel, add a small avatar circle with the participant's initials using the appropriate sentiment border colour. The live transcript text on /app/live should use a slightly different background (darker) to look like a terminal/console output.
---
## Final Touch Prompts
### Prompt F1 — Demo Mode Indicator
Add a subtle "Demo Mode" indicator badge in the top-right corner of every app page (not the landing page). It should be a small pill badge with an indigo outline and the text "Demo Mode" in light gray. It should be fixed-position so it stays visible while scrolling.
### Prompt F2 — Toast Notifications
When any non-functional button is clicked (like "View Campaign", "Export Report", filter dropdowns that don't work, etc.), show a small toast notification that slides in from the bottom-right corner saying "This feature is available in the full version." The toast should auto-dismiss after 3 seconds and have a dark background with white text and a subtle indigo left border.
### Prompt F3 — Interview Audio Waveform
On the Live Interviews page, add a small animated audio waveform visualisation next to each active interview — simple CSS bars that animate up and down to suggest audio activity. Use indigo colour for the bars. This should be purely decorative (not real audio) but adds to the sense that real interviews are happening.
---
## Dummy Data Summary
### Demo Brand
- **Company**: Kindred — a direct-to-consumer meal kit delivery company
- **Plan**: Growth Plan
- **Logo initial**: "K" in an indigo circle
### Research Studies
| Study | Status | Interviews | Type |
|-------|--------|------------|------|
| Why Customers Cancel After Month 3 | Completed | 80/80 | Exploratory |
| Competitor Switching Motivations | Completed | 80/80 | Strategic |
| Premium Tier Willingness to Pay | In Field | 48/75 | Evaluative |
| Eco Packaging Perception Study | Recruiting | 32/60 | Evaluative |
### Participants (20 total)
| Name | Age | Location | Study |
|------|-----|----------|-------|
| Sophie A. | 34 | Bristol | Churn Drivers |
| James T. | 44 | Manchester | Churn Drivers |
| Rachel H. | 29 | London | Churn Drivers |
| Olivia M. | 31 | Glasgow | Churn Drivers |
| Grace W. | 33 | Brighton | Churn Drivers |
| David L. | 41 | Edinburgh | Churn Drivers |
| Tom O. | 36 | Leeds | Churn Drivers |
| Noah K. | 27 | Cardiff | Churn Drivers |
| Daniel F. | 39 | Newcastle | Churn Drivers |
| Priya S. | 38 | Birmingham | Churn Drivers |
| Hannah B. | 32 | Dublin | Churn Drivers |
| Marcus O. | 28 | London | Eco Packaging |
| Kenji N. | 52 | Bath | Premium Tier |
| Maya W. | 37 | Edinburgh | Premium Tier |
| Aisha K. | 29 | London | Premium Tier |
| Ryan M. | 36 | Dublin | Eco Packaging |
| Lauren B. | 31 | Cardiff | Eco Packaging |
| Oliver R. | 42 | Liverpool | Premium Tier |
| Freya C. | 26 | Bristol | Eco Packaging |
| Nadia P. | 33 | Birmingham | Eco Packaging |
### Colour Palette
| Purpose | Colour | Hex |
|---------|--------|-----|
| Background | Near-black | #09090F |
| Card surface | Dark charcoal | #13131A |
| Card border | Subtle dark | #1C1C28 |
| Primary accent | Indigo/Violet | #7C5CFC |
| Primary gradient end | Light violet | #A78BFA |
| Positive / Success | Teal | #14B8A6 |
| Mixed / Neutral | Amber | #F59E0B |
| Negative / Warning | Rose | #F43F5E |
| Info | Blue | #3B82F6 |
| Heading text | White | #F5F5F7 |
| Secondary text | Light gray | #A1A1AA |