CareerPooler: AI-Powered Metaphorical Pool Simulation Improves Experience and Outcomes in Career Exploration

TL;DR

Qualitative findings show that spatial-narrative interaction fosters experience-based learning, resilience through setbacks, and reduced psychological burden in CareerPooler, a generative AI-powered system that employs a pool-table metaphor to simulate career development as a spatial and narrative interaction.

Abstract

Career exploration is uncertain, requiring decisions with limited information and unpredictable outcomes. While generative AI offers new opportunities for career guidance, most systems rely on linear chat interfaces that produce overly comprehensive and idealized suggestions, overlooking the non-linear and effortful nature of real-world trajectories. We present CareerPooler, a generative AI-powered system that employs a pool-table metaphor to simulate career development as a spatial and narrative interaction. Users strike balls representing milestones, skills, and random events, where hints, collisions, and rebounds embody decision-making under uncertainty. In a within-subjects study with 24 participants, CareerPooler significantly improved engagement, information gain, satisfaction, and career clarity compared to a chatbot baseline. Qualitative findings show that spatial-narrative interaction fosters experience-based learning, resilience through setbacks, and reduced psychological burden. Our findings contribute to the design of AI-assisted career exploration systems and more broadly suggest that visually grounded analogical interactions can make generative systems engaging and satisfying.

Figures (5)

• Figure 1: The main interface layout and core functional modules of CareerPooler. Including Game Mechanism Elements (left): Random event ball, Time cost line, Skill ball, Milestone Ball, and Pocket. And Information Display Modules (right): Achieved/All milestone, Time Information, Timeline and Events detail within the timeline.
• Figure 2: Workflow of System Interaction and Event Generation: After providing personal background and a two-year career goal, users enter the interactive interface. Users can hover over skill or random event balls to preview information hints, or pocket balls to collect events. Collected events include: skill balls, milestone balls, and random event balls. The simulation ends when either six milestones are collected or 730 days have elapsed, at which point a career journey report is generated.
• Figure 3: Study workflow: participants completed five phases: (1) study preparation with randomized system assignment; (2) pre-study survey on career clarity; (3-4) two career exploration tasks using CareerPooler and ChatGPT in counterbalanced order, each with post-surveys; and (5) semi-structured interview on experiences and AI's career impact. Total session time averaged 105 minutes.
• Figure 4: User journey of P11 over a 730-day simulation in CareerPooler. The participant began by providing a self-description and a career aspiration (“to become a researcher/professor specializing in machine learning and large language models”). In the early phase, P11 developed mentoring and social skills while struggling with initial student training and lab management. Midway, P11 obtained an assistant professor position, established a small research lab, and faced challenges such as attracting students, competing with a renowned neighboring lab, and securing limited funding. In subsequent milestones, P11 guided the first student cohort to complete projects, learned to design experiments more effectively, and refined project selection strategies despite technical setbacks and student attrition. In the final stage, P11 began preparing and revising research papers, marking the transition toward scholarly productivity. The journey concluded with a personalized career report summarizing milestones achieved, random events encountered, and skills acquired, reflecting both the accomplishments and difficulties of early academic career development.
• Figure 5: This figure presents the performance progression across different ball types and gaming sessions. For White Ball Shots per Session, participants showed improvement from the first session (M=30.7) to the most experienced players (3+ sessions, M=50.0). Balls Pocketed per Session increased from M=8.6 in the first session to M=13.6 for participants with three or more sessions. The ball type analysis shows progression across all three categories. Milestone Balls increased from M=3.2 (1st session) to M=4.9 (3+ sessions). Random Event Balls demonstrated improvement from M=2.3 to M=4.7 across sessions. Skill Balls showed growth from M=3.1 to M=4.0 across multiple gaming sessions.