Digital Library — UX Case Study

Kepler

atomic theory

Library

Results for "atomic theory"

4 results · AI-ranked ✦ AI All Years English Physics

Atomic Women

Roseanne Montillo · 2020

The untold story of the women scientists behind the atomic age and their remarkable contributions...

Atomic Bomb

How to Make an Atomic Bomb

Bob Dale · 2018

A historical account of the scientific breakthroughs that led to the development of nuclear weapons...

Atomic Physics

Atomic Physics for Everyone

R.L. Rhodes · 2020

An accessible introduction to the principles of atomic physics for students and general readers...

Making of

The Making of the Atomic Bomb

R.L. Rhodes · 2019

Pulitzer Prize-winning account of the Manhattan Project and the scientists who built the first atomic bomb...

01 — Brief

What the Digital
Library Does

An academic digital library is meant to provide authentication, source management, and access to restricted content. In practice — the traditional interface burdens both types of users and leaves them frustrated.

Research Question: What barriers do library users experience, and what actions can new design take to resolve them?

Research was conducted via structured survey distributed to 24 users — from first-year students to faculty. Each participant was asked about usage frequency, academic confidence, main pain point, and desired feature.

Participants

3.8

Tech Comfort

3.0

Academic Confidence

Academic Status Breakdown

Undergrad — 63% (15)

PhD — 17% (4)

Faculty — 12% (3)

Masters — 8% (2)

Pain Points

Filter overload & confusion9 / 24

Hard to verify recency6 / 24

No save & organize ability5 / 24

Outdated interface / too many clicks4 / 24

Usage Frequency

Only during assignment periods9 / 24

1–2 times a month6 / 24

Daily5 / 24

2–3 times a week4 / 24

What We Found

Finding 01

Majority: Low-Confidence Students

63% undergraduates, academic confidence 2.0/5. The dominant pain — filter overload and complex filtering.

Finding 02

Active Minority: Daily Researchers

37% advanced researchers, confidence 4.7/5. Missing personal workspace — each login starts from zero.

Finding 03

Recency — An Unsolved Question

6 of 24 cannot identify if an article is recent. Solution: prominent visual emphasis of publication year on every card.

Finding 04

Requested Features

Simple search, personal folders, citation export, year filtering — the four most-mentioned features.

02 — Research

Two Personas
Derived from Data

Not invented — built on the distribution the data dictated. Two groups with entirely different needs.

Undergraduate · Social Sciences · Age 24

Noa Cohen

"I just want one search field that gets me what I need — without losing an hour in filters I don't understand."

Metrics

Tech Skills4 / 5

Academic Confidence2 / 5

Usage FrequencySeasonal

Group63% of users

Pain Points

Result overload, no idea where to start

Complex, intimidating filters

Can't easily tell if article is recent

Interface feels "too professional"

Goals

One good article, fast and effortless

Simple search — one field, instant results

Easily spot date and relevance

Be guided step by step

PhD Researcher · Sociology · Age 34

Eitan Levi

"I need a place to manage all my research — not just search for articles, but save, organize by project, and export citations fast."

Metrics

Tech Skills3 / 5

Academic Confidence5 / 5

Usage FrequencyDaily

Group37% of users

Pain Points

No workspace — each login starts fresh

Outdated interface, too many clicks

Citation export is tedious

No project-based organization

Goals

Workspace for each research project

Citation export in one click — APA / MLA

Quick access to saved items

Integrated bibliography management

Direct Comparison

Attribute	Noa	Eitan
Average age (group)	23.2	37.8
Academic Confidence	2.0 / 5	4.7 / 5
Tech Skills	4.3 / 5	3.0 / 5
Primary Pain Point	Filter Overload	No Workspace
Core Need	Simplicity + Guidance	Management + Speed

03 — User Flow

User Journeys

Two entirely different paths — one searches to succeed quickly, the other comes to work.

User Flow A — Noa Cohen

Search for seminar paper · Seasonal access · Tight deadline

Enter Site

Deadline approaching

Type in Field

"social psychology"

2,847 Results

Immediate overload

Attempt Filter

Confusion with options

Filter by Year

Manual entry

Open Article

Check date

Download PDF

Success

Emotional Arc

Entry

Overwhelmed

Confused

Recovering

Satisfied

Success

Friction Point: Results page — 2,847 results with no guidance cause a sharp emotional drop.

Opportunity: Auto-filter (last 5 years as default) + publication year prominently featured on every card.

User Flow B — Eitan Levi

Daily research management · Focused search · Bibliography export

Daily Login

Authentication

Workspace

Check saved items

Focused Search

Specific terms

Advanced Filter

Year + journal

Save to Folder

"Chapter 3"

Export Citation

APA · One click

Back to Writing

Done

Emotional Arc

Entry

Workspace

Filter

Save

Export

Complete

Friction Point: No workspace — every login requires starting fresh. Citation export also spans multiple pages.

Opportunity: Personal homepage with recent items, folders, and direct citation export button from workspace.

04 — Site Map

The Site Architecture

Information architecture derived from the personas — a direct path for Noa, a deep workspace for Eitan, and one article card that serves both.

Home

Search Results

Full Article Page

Advanced Search

My Library

My Folders

Saved Articles

Reading List

Search History

Browse

By Subject

By Date

By Journal

Article Card

Download PDF

Export Citation

Citation Formats

Save to Library

Help

Search Guide

FAQ

Noa's path — Home → Search → Results → Article → PDF

Eitan's path — Home → My Library → Folders → Export Citation

Shared — Article Card: PDF for Noa · Citation for Eitan

Noa's Path

Home → Search → Results → Article → PDF. Short, direct, no branching.

Eitan's Path

Home → My Library → Folders → Export. Administrative path, workspace is home.

Shared Touch Point

Article Card — PDF for Noa, Citation for Eitan. Layered design, not duplicated.

05 — AI Copilot

The Digital Librarian —
AI Copilot

No more search engine. A conversational interface that knows the user, clarifies the question before showing results, and serves as the command center for the entire site.

Personalized Entry Experience

When a user logs in — the interface recognizes them. Dynamic placeholder, chips of recent searches, and instant access to what worked last time.

Conversational Filtering — The 4-Options Flow

When a user asks a broad question, the AI doesn't dump 2,847 results. It negotiates briefly — 4 clear options, with an escape hatch always available.

System Control — Omnipotent Assistant

The chat isn't just a search engine — it's the "command center" of the whole site. Help, navigation, explanations — all through natural conversation.

Interactive Help

"How do I export to APA?" — The chat not only explains, it brings the relevant button directly into the conversation. No menu hunting.

Smart Navigation

"Take me to my PhD research folder" — The chat navigates the user directly to the right place, no three-menu chain required.

Critical Design Decisions

Decision 01

Results in Interface, Not Chat

The chat is the funnel — it clarifies the question. But results pour into the main interface. This prevents the chat from becoming cluttered and cramped, keeping control in the user's hands.

Decision 02

Accessibility in Interface, Not AI

If a user can't see well, they won't find the chat to ask it to enlarge text. Accessibility must be a fixed, visible button in the header, always accessible, not AI-dependent.

Decision 03

Max 4 Options + Escape Routes

The 4-Options Flow works because it's bounded choice — not open-ended. But without escape routes ("skip", "type again", "X"), the chat becomes a blocker. With escape routes — it becomes an optional guidance path.

Decision 04

Voice Input = Accessibility, Not Core

Microphone icon in the input field — important accessibility support for specific use cases. But in a quiet academic setting, most use is typing. Nice-to-have feature, not MVP.

06 — Wireframes

Key Screens —
The Proposed Solution

Five core screens for both personas — every design decision derived directly from the pain points surfaced in the research.

Homepage

AI Search, smart defaults, 4-Options Flow

Search Results

Year badge, direct PDF, save to library

Full Article

PDF for Noa, APA citation for Eitan, folder

My Library

Folders, batch export, citations hub

Active Reading

Focus Mode, floating citation, Workspace sync

07 — Mockups

Actionable Design —
Three Core Screens

What the user actually sees. Every screen is built on one principle: fewer options, more power.

Search Results Screen

Screen 1 — Search Results

Results sorted by relevance, year badge prominent on every row, filters small and unobtrusive.

✓

Year badge on every row — badge color fades with article age. Recency is impossible to miss.

✓

Direct CTA per row — "Download PDF" + "Save" without entering the article page. Noa finds, downloads, moves on.

Eitan Levi · Lecturer / Researcher

Screen 2 — My Library

Organized workspace with folders, citation export, and sources by project.

✓

Folder organization by project — Eitan manages 4 parallel projects. Clear separation without manual syncing.

✓

Batch citation export — select multiple articles, export in one format. Solves Eitan's biggest pain point.

08 — Prototype

Interactive Prototype —
Try It Yourself

The full research translated into a clickable product. Navigate the search results, open a book, save it to your library, and read inline — the complete user journey in one prototype.

kepler.app / prototype

The interactive prototype is best experienced on desktop.
Tap below to open it in a new tab.

Open Prototype ↗

Step 1

Search Results

Browse the AI-curated results list

Step 2

Book Detail

Read preview & copy citation

Step 3

Personal Library

Organize saves by project folder

Step 4

Inline Reading

Highlight text & annotate inline

09 — Matrix

What to Build First —
Impact vs. Effort

Each feature plotted by research citation frequency (impact) versus relative development effort (effort). The top-right quadrant is the MVP.

MVP — Build Now

Single search field, year badge, smart defaults (last 5 years), direct PDF inline — four features, all derived from the 24 user pain points. Low dev effort, maximum immediate impact.

Phase 2 — After MVP

Batch citation export, personal library folders, and related articles. High impact for PhD students and faculty, but require backend infrastructure and auth.

10 — Conclusions

What We Learned

The research with 24 users points to four clear design principles. Each one translates directly into a feature in the new interface.

Principle 01

Fewer Options, More Power

37% cited "filter overload" as their primary pain point. The solution isn't to remove filters — it's to choose a default that works: last 5 years. Users arrive at good results without setting anything.

Principle 02

Recency Instantly Visible

25% struggled to identify if an article is recent. The year badge on every row solves it without the user hunting — it's right there, impossible to miss.

Principle 03

Workspace for Researchers, Search for Students

Both personas land on the same article page from completely different places. Layered design — PDF front and center, citation in the background — lets each person grab what they need.

Principle 04

MVP First, Workspace Later

The MVP — search, year badge, smart default, direct PDF — serves students who are 63% of users. The personal library arrives in Phase 2 for PhD students and faculty.

"Good design is when the user doesn't feel like someone designed something — they just find what they need."

24 users · 4 core pain points · Clear path from research to solution

Digital LibraryThrough the User's Eyes

What the DigitalLibrary Does

Academic Status Breakdown

Pain Points

Usage Frequency

What We Found

Two PersonasDerived from Data

Direct Comparison

User Journeys

The Site Architecture

The Digital Librarian —AI Copilot

Personalized Entry Experience

Conversational Filtering — The 4-Options Flow

System Control — Omnipotent Assistant

Critical Design Decisions

Key Screens —The Proposed Solution

Actionable Design —Three Core Screens

Interactive Prototype —Try It Yourself

What to Build First —Impact vs. Effort

What We Learned

Digital Library
Through the User's Eyes

What the Digital
Library Does

Two Personas
Derived from Data

The Digital Librarian —
AI Copilot

Key Screens —
The Proposed Solution

Actionable Design —
Three Core Screens

Interactive Prototype —
Try It Yourself

What to Build First —
Impact vs. Effort