PathFinder Errand Planner
Embedded Files
Project Overview
PathFinder is an AI-enhanced errand planning experience designed for busy adults living in walkable cities like New York. The project explores how artificial intelligence can reduce mental overload and decision fatigue by helping users plan errands in ways that feel adaptive, emotionally supportive, and human-centered rather than purely optimized for speed.
The project emerged from a simple but relatable tension: errands are rarely difficult on their own, but the mental effort surrounding them can feel exhausting. People constantly make small decisions about timing, route order, physical energy, transportation, and unexpected disruptions. Existing tools like Google Maps optimize efficiency, but they often ignore the emotional experience of navigating a city after a long day.
Audience: Busy adults in walkable cities who batch tasks after work and crave both efficiency and small rewards.
Duration: 7 Weeks
Team: Rutuja, Grace & Sofia
Role of AI: Personalization, prediction, and generative reasoning, working quietly in the background.
Output: A live interactive prototype built with vibe coding, featuring real Claude API integration for task scoring.
The Problem
Running errands is often framed as a logistical task, but our research revealed that the real issue is cognitive and emotional fatigue.
Many users described errands as a “second shift” after work. Even simple tasks became overwhelming because of the constant micro-decisions involved: deciding route order, checking store hours, carrying heavy items, avoiding crowds, and adapting when plans unexpectedly changed. While people cared deeply about efficiency, they also wanted moments of ease and enjoyment woven into the experience.
This became the central design opportunity for PathFinder. We were not trying to eliminate errands entirely. Instead, we wanted to reduce the emotional friction surrounding them.
Research & Key Insights
Our process began with desk research, interviews, and exploratory persona building. The early research phase focused on understanding how people currently organize errands and how they emotionally experience those routines.
Several patterns emerged consistently across our findings.
➡️ Users already mentally batch errands by geography and urgency. Many described building “loops” around the city to avoid backtracking. However, current tools mainly support speed optimization rather than emotional comfort or adaptability.
➡️ We also discovered that users rarely think of AI as something useful for everyday planning. Most participants associated AI with large or complex tasks like vacation itineraries rather than ordinary routines. This created an important design challenge: if we wanted people to trust AI in a daily context, the system needed to feel understandable, lightweight, and genuinely helpful.
➡️ Another major insight was that delight mattered more than we initially expected. While efficiency was necessary, participants repeatedly described moments of joy that transformed errands from chores into meaningful experiences. Unexpected discoveries like coffee shops, bookstores, parks, or pleasant walking routes made the process feel more worthwhile.
Our interviews also revealed frustrations that existing tools largely ignore. Users spoke about physical exhaustion, heavy carrying loads, unpredictable wait times, crowded environments, and decision fatigue after work. One participant explained that even small choices became mentally draining after a long day.
These findings pushed the project beyond pure navigation design. Instead of designing only for optimization, we began designing for emotional pacing, adaptability, and intentional living.
Interactive Personas
To ground the project in realistic behavior, we developed two primary personas:
Alex and Marcus.
Alex: The Overwhelmed Young Professional
Alex's persona directly influenced our explainability decisions. We realized the AI could not simply generate routes invisibly. Users needed to understand why the system made decisions.
Marcus's helped us think about adaptability over time. Unlike Alex, he was already willing to experiment with AI. His needs centered more around flexibility, route adjustment, and minimizing wasted effort.
Together, these personas revealed a shared emotional tension that people wanted support without feeling controlled.
Early Explorations & Design Direction
Mood Boards
Before building the final prototype, we explored several emotional and visual directions through AI-generated Mood Boards, Journey Mapping and Concept Sketches.
The earliest explorations focused on ideas like “Calm Efficiency,” “Confident Clarity,” and “Effortless Living.” Across these concepts, we consistently gravitated toward soft color palettes, natural textures, open spacing, and low-friction interactions.
The selected direction, Effortless Living, uses a brighter but still muted palette that keeps users calm and motivated. We intentionally avoided aesthetics commonly associated with hyper-productivity apps. Instead of bright notifications and aggressive optimization language, we wanted the interface to feel calm, lightweight, and emotionally breathable.
This direction later evolved into the final PathFinder visual system, which used muted greens, warm neutrals, and nature-inspired imagery to reinforce intentionality and reduced cognitive load.
Adaptive Journey Mapping
How the AI experience evolves over time
As the visuals evolved, we began thinking about PathFinder not as a one-time planner but as an adaptive relationship between user and system over time.
Our Adaptive Journey Map explored how the AI gradually learns from behavior, preferences, and emotional patterns.
At first, the AI behaves conservatively, focusing mostly on practical route optimization. Over time, it becomes more personalized by recognizing things like preferred walking environments, tolerance for crowds, or recurring patterns in energy levels.
This evolution was important because we wanted the AI to feel assistive rather than invasive. The system slowly adapts through interaction instead of assuming it already “knows” the user.
The Core AI Concept
PathFinder combines three major AI-enhanced systems:
Energy Budgeter
Buffer Guard
Serendipity Guide
Together, these features transform errands from static checklists into adaptive experiences.
Incorporation of the 3 Concepts
Concept 1: Energy Budgeter
The Energy Budgeter became the emotional core of the project. Inspired by our research around burnout and decision fatigue, this feature allows users to rate their energy on a 1–10 scale.
The AI then dynamically reorganizes and filters the errand list based on that emotional state.
When energy is low, the system “prunes” the list to prioritize only essential tasks while minimizing walking distance and physical strain. Higher energy levels unlock more ambitious routes or optional stops.
This feature intentionally reframed productivity. Rather than encouraging users to maximize output, PathFinder adapts expectations to the user’s current capacity.
The system also changes route behavior based on energy levels. Low-energy paths prioritize quieter streets, fewer stairs, flatter terrain, and shaded walking routes.
This became an important emotional design decision because many navigation systems assume users always want the fastest route. Our research suggested otherwise. Sometimes users wanted the least stressful route.
Concept 2: Buffer Guard
Buffer Guard was designed to respond to real-world unpredictability. Participants consistently described frustration when one delayed stop disrupted their entire evening.
To address this, Buffer Guard monitors factors like store hours, crowd levels, and estimated wait times to proactively suggest route changes.
For example, if Trader Joe’s suddenly becomes crowded, the app recommends swapping the grocery stop with another nearby errand instead of forcing the user into an unnecessary wait. The system explains why the suggestion was made rather than silently changing the route.
This feature addressed explainability and trust directly. Instead of feeling like a black-box AI system, PathFinder behaves more like a collaborative planning assistant.
Concept 3: Serendipity Guide
One of the most unique aspects of PathFinder was the Serendipity Guide. While most planning tools focus entirely on efficiency, our research revealed that people emotionally valued small moments of discovery.
The Serendipity Guide introduces optional “micro-joy” stops along a user’s route, such as cafés, bookstores, murals, gardens, or scenic walking paths.
Importantly, these moments remain fully dismissible. Users can skip them, permanently remove certain suggestions, or rate whether they enjoyed them. These ratings persist across sessions and help the AI personalize future recommendations.
This feature represented a broader philosophical shift within the project. We stopped thinking about errands as something users simply needed to “finish” and instead explored how routine urban experiences could become more emotionally rewarding.
Building the Interactive Prototype
The final prototype expanded significantly beyond the original concept boards.
The app includes three main tabs: Overview, Route, and Map. Users can add errands, adjust their energy level, review AI-generated route suggestions, and navigate through the city using adaptive walking guidance.
The Overview tab functions as the planning center. Users add tasks through a location autocomplete system, and the AI immediately prioritizes errands based on urgency, timing, and user energy.
The Route tab introduces different “vibes” for navigation: Efficient, Low-Friction, and Scenic. Each route explicitly communicates what the AI optimized for, helping users understand the reasoning behind recommendations.
The Map tab visualizes the full route and includes optional joy stops woven naturally into the experience. During navigation, users can visit, skip, or permanently dismiss suggestions. After completing a route, users rate their experience, allowing the AI to improve future recommendations.
The prototype also integrated lightweight AI behaviors using Claude for urgency scoring and adaptive recommendations.
AI Ethics, Trust & Human-Centered Design
A major focus of the project was ensuring the AI remained supportive rather than overly controlling.
Throughout the process, we consistently asked ourselves:
How can AI reduce mental effort without removing user autonomy?
Several design decisions emerged from this question.
First, we emphasized explainability throughout the interface. The app clearly communicates why it suggests certain routes, why a stop was reordered, or why a detour appears. This directly addressed user distrust uncovered during interviews.
Second, we designed for optionality rather than automation. Joy stops are dismissible. Users choose route styles intentionally. The AI suggests rather than commands.
Third, the system supports intentional living rather than engagement maximization. Earlier concept explorations explicitly described designing an app that “knows when to step back.” This philosophy shaped the overall tone of the experience.
We wanted PathFinder to help users finish errands efficiently and then return to life beyond the screen.
Key Learnings
One of the biggest lessons from this project was realizing that efficiency alone does not create meaningful experiences.
Initially, we approached the project primarily as a route optimization problem. However, our research quickly revealed that emotional experience mattered just as much as logistical efficiency. People were not simply trying to save time; they were trying to reduce stress, regain energy, and feel better about how they spent their day.
We also learned that trust is essential in AI-enhanced systems. Users were far more receptive to AI when the system communicated clearly and respected their autonomy.
Another important takeaway was that AI can support emotional well-being in subtle ways. The most successful ideas in the project were often the smallest ones: suggesting a quieter walking route, delaying a stressful stop, or introducing a pleasant detour at the right moment.
Rather than positioning AI as a hyper-intelligent decision maker, PathFinder reframed AI as a gentle adaptive companion.
Reflection & Next Steps
If we continued developing PathFinder, we would expand real-time environmental awareness by integrating live transit conditions, weather data, and crowd density. We would also explore accessibility features more deeply, including mobility-friendly route customization and sensory-friendly navigation options.
Future testing would also help evaluate whether the emotional pacing of the system genuinely reduces stress over longer periods of use.
Most importantly, this project shifted our understanding of what AI design can be. Instead of focusing purely on automation and optimization, we explored how AI might support more intentional, emotionally aware everyday experiences.
PathFinder ultimately became less about errands and more about helping people move through their lives with less friction and more ease.
Page updated
Google Sites
Report abuse