Floofers: transforming pet care with AI intelligence
Making pet-sitting decisions 3× faster with targeted AI — not an AI product, but a product that uses AI where it matters.
Projected impact
Pre-launch projections modeled from Floofers' historical funnel data (verification → search → booking).
95%↓
verification time
3–5 days → 2 hours
65%↓
search duration
23 min → 8 min
171%↑
booking conversion
34% → 92%
Role
UX/UI Design Consultant
Timeline
3 months
Tools
Figma, Adobe Illustrator, Photoshop
Collaborators
Founding team, data scientist, engineering
Constraints
3-month timeline · 1 designer · existing engineering team
My role, specifically
Research synthesis · UX flows · UI design · AI microcopy · human-in-the-loop logic
Background
Floofers connects pet parents with vetted sitters across Australia. Three friction points were quietly costing growth and trust. I designed three targeted AI features to fix them.
The problem
The experience was slow and overwhelming in three specific ways.
01
Slow verification
3–5 day approval delay capped platform supply growth.
02
Choice overload
23 min average scrolling near-identical profiles before booking.
03
Generic profiles
Strong candidates couldn't surface relevant experience — generic bios don't get booked.
Business impact — baseline, pre-redesign
34%
booking abandonment
23 min
avg. decision time
3–5 days
verification delay
40+
daily support calls
Research & discovery
8 in-depth interviews + 40 survey responses. Stakeholder workshops to separate business assumptions from real user needs.
35% of owners
The Anxious First-Timer
Needs peace of mind — can't gauge trust from profiles alone.
"I read every review twice and still felt nervous."
45% of owners
The Busy Professional
Needs reliable care booked fast — too many similar profiles to compare.
"I don't have 20 minutes to compare sitters — I need the right one, fast."
20% of owners
The Special-Needs Pet Parent
Needs specialized experience surfaced — generic bios bury it.
"I need someone who's handled insulin shots."
What stakeholder interviews surfaced
- — People expect the system to learn and adapt to their pet's preferences over time.
- — Speed and automation were essential for verification to scale with growth.
- — New sitters needed guidance to build a profile that actually gets booked.
- — Transparency was the deciding factor in whether either side trusted the platform.
- — A visible "why this match" reason built confidence faster than a rating alone.
From research to direction
Three friction points surfaced across every persona. Affinity mapping with the founding team confirmed alignment with business priorities.
Affinity Map — Interview & Survey Clusters
Verification Bottleneck
"Waiting 5 days kills my onboarding momentum as a new sitter."
"Applications fail with zero explanation on what went wrong."
Choice Overload
"Every sitter profile looks identical — 5 stars and generic bios."
"I scroll for 20 minutes and still feel unsure if they fit my pet."
Generic Profiles
"Sitters just write 'I love animals' instead of relevant skills."
"Specialized experience (e.g. senior dogs, medical) gets buried."
"Instead of adding AI everywhere, I focused on the three moments that mattered most to users."
Solution 01
Quick verification for pet sitters
Problem
Upload → wait 3–5 days → resubmit → wait again. Bottlenecked supply growth.
Solution
AI-powered verification with instant feedback. Human review for edge cases only.
Getting there — early direction & wireframe explorations
Key pivot: Fully automated pass/fail → sitters didn't trust black-box decisions. Pivoted to hybrid: AI scans clear cases, flags ambiguous ones for human review.
❌ Explored: 100% Automated Pass/Fail
Black-box decisions caused sitter anxiety. OCR misreads spiked support tickets.
✓ Adopted: Hybrid AI Scan + Human Review
Clean documents approved in under 2 hours. Ambiguous cases route to Certn.co human review.
Before
3–5 day manual review, opaque status
After
AI scan + 2-hour approval
Visual feedback
- — Uploading → scanning → result state progression
- — Consistent naming: "ID Verification" / "Police Check"
- — Specific errors, one action per step
Human + AI balance
- — AI handles clear approvals
- — Certn.co for accredited background checks
- — Human reviews edge cases
95%↓
verification time
3–5 days → 2 hours avg.
67%↓
fewer document
resubmissions
Solution 02
Smart search for pet owners
Problem
23 minutes scrolling generic profiles with no way to gauge fit.
Solution
Preference-based matching with plain-language reasons on every result.
Getting there — early direction & wireframe explorations
Key pivot: Raw percentage scores ("94% Match") felt arbitrary in testing. Replaced with plain-language reasons ("3 years with senior dogs") on every card.
❌ Explored: Raw % Match Score
Bare percentages created skepticism — owners couldn't verify the number.
✓ Adopted: Plain-Language Reasons
"Experienced with anxious pets" / "Live within 1 km" — verifiable at a glance.
Solution 02: User Flow Diagram
Owner Journey & AI Logic TouchpointsBefore
Generic list, 23 min decision time
After
Smart matching, 8 min to book
What the matching model weighs
Explaining AI decisions
Users needed to know why, not just who. Every result shows:
- Primary match reason
- Secondary factors (location, availability)
- One genuinely different alternative
65%↓
search time
23 min → 8 min
58%↑
search-to-booking
conversion, this feature alone
Solution 03
AI-assisted profile builder for sitters
Problem
Sitters didn't know what to highlight — generic profiles don't get booked.
Solution
Guided profile flow with AI writing assistance, optimized for what owners search for.
Getting there — early direction & wireframe explorations
Key pivot: Auto-published bios felt generic and inauthentic. Switched to "AI drafts, human edits and approves" — writing assistant, not decision maker.
❌ Explored: Auto-Published Bio
Generic AI text that sitters disowned. Low engagement, manual rewrites.
✓ Adopted: AI Drafts, Human Approves
Dynamic questions → editable first draft. Sitter keeps full ownership.
Before
Blank text fields, generic bios
After
AI-guided drafting, sitter approval
Step 1
Experience analysis
Adaptive questions surface specializations.
Step 2
Smart suggestions
Bio writing help + local market pricing guidance.
Step 3
Profile strength
Real-time meter with actionable tips, not just a %.
AI writing assistant in practice
71%↑
profile completion
52% → 89%
4 days
avg. time to first booking
down from 2 weeks
Technical implementation
Proven services wired together — my effort went into how results are shown and explained.
Data privacy & ethics
- ✓ Personal data encrypted and anonymized before training
- ✓ Clear, explicit consent for every AI-powered feature
- ✓ Opt-out available for any AI functionality
- ✓ Regular bias audits across pet type, price tier and region
Full funnel breakdown
Projected results, full funnel
(Same modeling basis as above — shown together here to illustrate compounding impact across the three features.)
95%↓
Verification time — 3–5 days → 2 hours
65%↓
Search duration — 23 min → 8 min
71%↑
Profile completion — 52% → 89%
171%↑
Booking conversion — 34% → 92%
73%↑
Repeat bookings — 45% → 78%
Search relevance lifted conversion 58% alone. Layered with faster verification and stronger profiles, platform-wide improvement compounds to 171%.
+42%
total platform bookings
$2.3M
additional revenue, 6 months post-launch
−29%
customer support volume
What I learned
01
Visible reasons > bare scores.
"3 years with senior dogs" builds more trust than "94% Match" ever could.
02
Speed without transparency feels reckless.
The fix wasn't slowing down — it was showing what's happening honestly.
03
AI should draft, never publish.
Every AI output is a first pass the human reviews.
04
The best AI feature is invisible.
Users said "I found the right sitter fast" — not "I love the AI."
Looking ahead
Next: seasonal recommendations, pet health data integration, predictive sitter availability. Same foundation — technology that helps humans decide better.