Sentience

Sentience

Case study developed as part of the Google UX Professional Certificate, applying real-world research constraints and industry-standard UX methods.

Case study developed as part of the Google UX Professional Certificate, applying real-world research constraints and industry-standard UX methods.

Year

Platform

Category

Duration

Project Link

2025

Web

B2c, Saas

6 Weeks

Overview

A standalone productivity workspace that learns from your behavior to predict what you need without reading your content.

A standalone productivity workspace that learns from your behavior to predict what you need without reading your content.

A standalone productivity workspace that learns from your behavior to predict what you need without reading your content.

Problem

It's 2:58 PM. Your 3 PM meeting starts in two minutes. You had notes on this somewhere, Notion? Google Docs? Slack? You open four tabs searching. You find it at 3:04 PM. Everyone's been waiting.

The issue isn't bad tools. It's that good tools don't talk to each other. You lose time finding things instead of using them.


Target User: Knowledge workers managing multiple projects across disconnected apps who lose time searching, context-switching, and manually organizing workflows that follow predictable patterns.

Deliverables:

Core Interface designs

  1. Morning dashboard showing contextual app surfacing

  1. Focus mode with minimal interface

  2. Evening wrap-up mode

  1. Morning dashboard showing contextual app

    surfacing

  1. Focus mode with minimal interface

  2. Evening wrap-up mode

Feature Demonstrations

  1. Task prediction scenario (end-to-end flow)

  2. Pattern learning progression

  3. Location-based UI adaptation (home/office/mobile contexts)

  1. Task prediction scenario (end-to-end flow)

  2. Pattern learning progression

  3. Location-based UI adaptation (home/office/mobile

    contexts)

System Details

  1. Character customization options (personality styles, visual themes)

  2. Onboarding flow for new users, understanding user preferences(Cold start handling)

  3. Settings and privacy controls

  4. Information architecture

  1. Character customization options (personality styles,

    visual themes)

  2. Onboarding flow for new users, understanding user

    preferences(Cold start handling)

  3. Settings and privacy controls

  4. Information architecture

Competitive Landscape

The ALL IN ONE'S - Notion, ClickUp, Coda

Centralize everything, but require migrating your entire workflow into their ecosystem. Limitation: High adoption friction, doesn’t work with tools users already rely on.

THE APP AGGREGATORS - Station, Rambox, Wavebox

Organize multiple apps in one window. Limitation: Still manual navigation, no intelligence about timing or context.

THE AI ASSISTANTS - Google Assistant, Siri, Alexa

Execute single commands well. Limitation: Reactive, not proactive. No pattern learning or workflow prediction.

THE AUTOMATIONS - Zapier, IFTTT, Apple Shortcuts

Enable custom workflows through rules. Limitation: High setup effort, nothing adapts automatically over time.

THE LAUNCHER APPS - Alfred, Raycast, Spotlight

Fast search and shortcuts. Limitation: Search-based, requires knowing what to look for.

THE SMART WORKSPACE ANALYTICS - Microsoft Viva, Workplace Analytics

Surface insights and behavioral patterns. Limitation: Reporting only, doesn’t act on insights in real time.

The gap Sentience addresses

The gap Sentience addresses

Cross-app intelligence

works with your existing tools instead of replacing them

Solution Window

Solution Window

Contextual Intelligence

Contextual Intelligence

Core features

Behavioral AI Dashboard

  • Task management with pattern recognition.

  • Calendar integration with time-blocking.

  • Today's Brief: Daily summary with AI-optimized work windows.

  • Contextual suggestions based on time, location, and activity.

Behavioral pattern detection (morning state)

Adaptive Modes

  • Focus Mode: Deep work state with minimal distractions.

  • Task Mode: Organizational view for planning.

  • Location-Aware: Adapts based on home/office/travel.

  • Time-Based: Morning planning, afternoon focus, evening wrap-up.

Behavioral pattern detection (morning state)

Behavioral AI Dashboard

Behavioral AI Dashboard

  • Task management with pattern recognition.

  • Calendar integration with time-blocking.

  • Today's Brief: Daily summary with AI-optimized work windows.

  • Contextual suggestions based on time, location, and activity.

Behavioral pattern detection (morning state)

Adaptive Modes

  • Focus Mode: Deep work state with minimal distractions.

  • Task Mode: Organizational view for planning.

  • Location-Aware: Adapts based on home/office/travel.

  • Time-Based: Morning planning, afternoon focus, evening wrap-up.

Behavioral pattern detection (morning state)

Predictive Surfacing

  • Your meeting starts in 18 minutes. Here's the doc you opened before the last 3 Project X meetings.

  • This task usually takes 40-60 minutes. I found an open window at 2:15 PM.

Behavioral pattern detection (morning state)

Predictive Surfacing

  • Your meeting starts in 18 minutes. Here's the doc you opened before the last 3 Project X meetings.

  • This task usually takes 40-60 minutes. I found an open window at 2:15 PM.

Behavioral pattern detection (morning state)

Character-Based Transparency

  • AI reasoning explained through a character interface.

  • Customizable tone (Minimal, Friendly, Professional, Playful).

  • Shows why predictions are made, building trust.

  • I noticed you typically do this after meetings, so I surfaced it.

Behavioral pattern detection (morning state)

Activity-Based Prompt Generation

  • Chat input dynamically changes based on context.

  • "Ask for meeting summary" appears before meetings.

  • "Draft presentation" surfaces during creative morning hours.

  • Progressive learning: generic Week 1 → personalized Week 4+.

Behavioral pattern detection (morning state)

Character-Based Transparency

  • AI reasoning explained through a character interface.

  • Customizable tone (Minimal, Friendly, Professional, Playful).

  • Shows why predictions are made, building trust.

  • I noticed you typically do this after meetings, so I surfaced it.

Behavioral pattern detection (morning state)

Activity-Based Prompt Generation

  • Chat input dynamically changes based on context.

  • "Ask for meeting summary" appears before meetings.

  • "Draft presentation" surfaces during creative morning hours.

  • Progressive learning: generic Week 1 → personalized Week 4+.

Behavioral pattern detection (morning state)

View All Screens here

View All Screens here

How Sentience Access your data

How Sentience Access your data

It only detects the patterns and analyses

1
1
1

What Makes Sentience Different?

Hybrid Architecture: Standalone + Intelligence Layer

  1. Complete productivity system (tasks, calendar, AI chat, focus modes)

  2. Works immediately without integrations

  3. Users can create tasks, schedule events, manage work independently

Intelligence Enhancement

  1. Connects to Gmail, Slack, Google Drive, Calendar, etc.

  2. Analyzes metadata only (filenames, timestamps, patterns)

  3. Learns behavioral patterns across apps

  4. Predicts what you need before you search

Key Differentiator

  • Metadata-only intelligence that's privacy-safe, content-blind, and behaviorally predictive.

Design Decisions

01

Adaptive UI based on context

The interface changes throughout the day based on detected patterns:

  1. Morning mode, it highlights planning essentials like tasks, today’s brief, and meeting summaries.

  2. Afternoon Mode, it highlights planning essentials like tasks, today’s brief, and meeting summaries.

  3. Evening Mode, it shifts into wrap-up mode, surfacing unfinished tasks and helping prepare for the next day.
    Problem it solves: Static dashboards show the same information regardless of what you’re doing. Sentience reduces cognitive load by matching the interface to the user’s current need.

02

Smart App Surfacing

Rather than showing all apps all the time, Sentience predicts what comes next based on observed patterns.
The prediction logic:
1.Pattern recognition: "You typically open docs and Slack after meetings"
2. Context awareness: "This meeting was about Project X".
3. Predictive surfacing: "Here's the Project X workspace and communication channel".
Problem it solves: Eliminates the "now where did I put that?" moment. The system anticipates next steps in your workflow instead of making you remember and navigate.

03

Onboarding & AI Characters

Sentience combines a character-based interface with personalized onboarding to build early trust. Users choose the AI’s tone, set preferences, and define how they want support.

Throughout the experience, the character explains why certain suggestions appear, turning complex AI logic into clear, human-friendly reasoning.

Problem it solves: New AI tools often feel opaque or intrusive. Transparent communication and user-defined tone make Sentience feel understandable, safe, and personal from the start.

04

Standalone-First Architecture

Sentience works as a complete productivity system from day one, giving users immediate value without requiring migrations or integrations. Intelligence improves gradually—from generic suggestions in week one to fully personalized insights as patterns build—while integrations remain optional enhancements. Problem it solves: Most tools demand setup or app connections before they’re useful. Sentience eliminates that barrier, ensuring users benefit instantly and adopt deeper features at their own pace.

Design Decisions

01
Adaptive UI Based on Context
The interface changes throughout the day based on detected patterns: 1. Morning mode, it highlights planning essentials like tasks, today’s brief, and meeting summaries. 2. Afternoon Mode, it highlights planning essentials like tasks, today’s brief, and meeting summaries. 3. Evening Mode, it shifts into wrap-up mode, surfacing unfinished tasks and helping prepare for the next day. Problem it solves: Static dashboards show the same information regardless of what you’re doing. Sentience reduces cognitive load by matching the interface to the user’s current need.
Adaptive UI Based on Context
02
Smart App Surfacing
03
Onboarding & AI Characters
04
Standalone-First Architecture

Design System

INFORMATION ARCHITECTURE

IA Hierarchy overview

Competitive Positioning

Sentience

Vs

Constraints & Decisions

Constraints & Decisions

Technical constraints

  1. Requires API access to multiple productivity apps

  2. Pattern recognition needs minimum 1-2 weeks of data

  3. Real-time context detection (location, focus mode) requires background processes

Scope decisions

  1. Started with 5 core apps (Calendar, Docs, Slack, Email, Tasks) rather than trying to integrate everything

  2. Focused on predictive surfacing rather than building new productivity features

  3. Kept character interactions minimal to avoid annoyance

Alternative Approaches Considered

Widget-based dashboard

Rejected because it's just another aggregator. Doesn't solve the core problem of knowing what to look at.

Silent AI automation

Rejected because users don't trust predictions they don't understand. Character explanation was necessary for adoption.

Manual rule-setting

Rejected because it defeats the purpose. Users don't want to configure "if meeting ends, show these apps" the system should learn that.

Key Learnings

Conclusion

  • Sentience explores whether predictive interfaces can reduce the cognitive overhead of managing multiple productivity tools. The core bet is that context-aware surfacing showing the right app at the right moment based on learned patterns is more valuable than comprehensive dashboards that display everything.

  • As a concept project, the biggest open question is whether users will grant the level of access required for pattern analysis, and whether prediction accuracy can reach the threshold where it saves more time than it costs in wrong suggestions.

  • The next step would be prototyping one high-fidelity scenario to test whether AI predictions feel helpful or intrusive, and whether character-based explanations provide genuine value or just add visual noise.

Sentience

Sentience

Next up  
Next up  

Create a free website with Framer, the website builder loved by startups, designers and agencies.