“UX ANALYSIS:
PERPLEXITY AI”
“Balancing User Expectations with AI Design Principles”
MOHAMMAD ARIF KHAN
Overview
What are AI Citations & Follow-Ups? 🔹 Why Analyze It?
Citations: References that link AI-generated answers to Trust & Credibility: Citations are a trust-building feature in AI. Without
original sources (articles, research papers, websites). They them, users may doubt responses.
help users verify information, ensuring transparency and Usability & Learnability: Follow-ups reduce friction in continuing a
trust. conversation and support novice users.
Follow-Ups: Suggested or user-initiated related questions Differentiation: Perplexity AI and similar products stand out because
after an initial query. They guide exploration and maintain they emphasize explainability and guided exploration.
engagement, reducing the cognitive effort of thinking
Product Impact: These features influence adoption, retention, and
about “what to ask next.”
satisfaction making them critical to evaluate from a UX perspective.
🔹 Use Cases
User Flow on Perpexility APP
Followup/share save
plan selection Type Query
Ai Generated answer
Open App onboarding/search
LAWS OF UX — APPLICATION & VIOLATION
Hick’s Law Fitts’s Law
Feature list is “Get Pro” button is
streamlined with bullets prominent and easy to Jakob’s Law
for quick scanning tap Top-right share icon
at onboarding is
→
unusual doesn’t
match expectation
Pareto Principle
All benefits are equally emphasized —
the top value props (e.g., “10x
citations”) should stand out
Hick’s Law
Minimal choices upfront
(just search bar + one
Law of Proximity
Mental Model promo card) reduces Search input + mic
decision friction grouped tightly at
Monthly/Yearly
toggle and bottom
→
bottom easy to see as
one interaction
CTA feel familiar
LAWS OF UX — APPLICATION & VIOLATION
Jakob’s Law
Search layout, top bar icons
(bookmark, share) match
patterns from other app Law of Proximity
Citation numbers float
without clear grouping
(tiny, scattered)
Law of Proximity
Question title + answer
grouped visually together
Cognitive Load
Fitts’s Law Clean typography +
spacing in text reduces
Large bottom “Ask follow- reading strain
up” bar is easy to tap
LAWS OF UX — APPLICATION & VIOLATION
Selective Attention
Fitts’s Law
Bold text + headings
Small superscript citations (¹²³⁶)
near text body are harder to tap on grab attention well
mobile
Von Restorff Effect
Jakob’s Law
Common icons for share, Related section links look
copy, and headphones at visually identical, none
bottom match patterns stand out to drive click
users know
Hick’s Law
Mental Model
Minimal visible options: main text +
“Ask follow-up…” field at expandable “Related” accordions keeps
bottom matches decisions light
chat/messaging app model
NIELSEN'S 10 USABILITY HEURISTICS
KEY TAKEAWAYS & SUGGESTIONS
FINAL UX SCORECARD
UX Area Score
learnability 3
Trust & Feedback 4
UX AREA
efficiency 5
Aesthetic Design 4
Error Management 3
0 1 2 3 4 5
UX AREA SCORE
Suggested UX Improvements:
FUTURE WIREFRAMES
→
1. Precise citation linking highlight exact
sentence/paragraph in source
→
2. Tooltip onboarding e.g., “Tap citation
to view source details”
3. Clarify Pro/Max features → add short
descriptors or info icons inline
4. Label key icons (bottom nav, bookmark,
share) to reduce recall load
→
5. Contextual help microcopy, examples,
or “?” info buttons in subscription & home
screens
THANK YOU
FOLLOW ME FOR MORE SUCH
CASE STUDY
Mohammad Arif Khan
