Reimagining the Student Work Experience at Tufts’ Tisch Library
OVERVIEW
At Tufts University’s Tisch Library, student workers are the first line of support for thousands of library visitors. But during high-traffic times—especially at the start of the semester—their workflow becomes cognitively draining, repetitive, and inefficient.
As part of my Human Factors Engineering course, I worked on a semester-long research and design project to identify and solve these ergonomic issues. Our final solution? A human-centered AI-powered chatbot that streamlines repetitive inquiries and frees up mental bandwidth for student workers.
Occupational Ergonomics | Human-Factors Engineering
This project was completed as part of my Methods for Human Factors Engineering class at Tufts Human Factors Engineering Department
MY ROLE
Human-Factors Researcher
UX Research, Cognitive Ergonomics, Literature Review, AI Tool Design, NASA-TLX, Stakeholder Interviews
DURATION
Fall 2024 semester
TOOLS
Google Forms, ChatGPT API, NASA-TLX, Qualitative Interviews
TEAM
4 members
Part 1: Identifying the Challenge
We began with field observations and interviews with student employees to understand the daily challenges they face at the service desk.
KEY FINDINGS
Repetitive Cognitive Load: 80% of staff time during peak weeks is spent answering FAQs like “Where are the printers?” or “What are the library hours?”
New Hire Strain: New workers frequently interrupt senior staff for help with basic inquiries.
Physical Fatigue: Long sedentary shifts and harsh lighting led to eye strain and back pain.
Mental Fatigue: Staff described their job as “mentally numbing” during slow periods and “overwhelming” during busy ones.
Part 2: Reviewing the Literature & Proposing Solutions
We validated our findings with academic research in ergonomics, AI, and cognitive workload.
LITERATURE REVIEW
Cognitive Ergonomics: Repetitive tasks are a major contributor to mental fatigue (Yadav, 2023)
Generative AI: Increases productivity by 15–30% and reduces customer frustration (Brynjolfsson et al., 2024)
NASA-TLX: We used this validated metric to assess perceived workload and design our intervention around lowering cognitive demand.
PROPOSED INTERVENTION
We proposed a custom AI-powered Chatbot, trained on Tisch Library FAQs and integrated into the existing digital infrastructure via QR codes and the “Let’s Chat” feature. This chatbot would:
Handle repetitive questions instantly
Support new hires as a learning companion
Free up time for staff to focus on complex tasks
DESIGN GOAL
Lower cognitive load, enhance worker autonomy, and improve student service quality.
Part 3: Prototyping & Recommendations
We developed a prototype using OpenAI's GPT-4 model customized as the "Tisch Library Virtual Assistant." Our testing showed:
4-second response time (vs. 15s for humans)
95% accuracy on basic FAQs
Lower mental load reported in NASA-TLX simulations
Positive reception from students and staff
IMPLEMENTATION PLAN HIGHLIGHTS:
QR codes across the library
Training modules for staff
Feedback loop to continuously improve chatbot performance
FUTURE EXPANSIONS:
Interactive library maps
Tablet-based access points
Deeper integration with the Tisch Library database
VIEW FINAL REPORT :
TRY OUT THE AI CHAT BOT!