Upskilling with AI with Key Lime Interactive

Upskilling with AI with Key Lime Interactive

Students are trying to find jobs. Early career professionals are trying to keep theirs. AI has made it harder for both

Project type

Exploratory Research

duration

4 Months

YEAR

2025

the Problem

Students and new professionals are lost in their upskilling journeys, rendering their efforts ineffective and misaligned with the practical skills demanded by the industry.

Learners are frequently lost in an overwhelming volume of resources, making their skill development efforts directionless and inconsistent. This results in a critical gap between theoretical knowledge and the practical skills demanded by the industry, leaving new talent unprepared for career opportunities.

5Million
5Million

Americans search for a job every month

The Context

What is Key Lime Interactive?

Key Lime is a user experience service design company that assists clients across multiple different industries. It is a women and minority owned company whose mission is to take a human-centered approach to help clients build improved services and products

Why are they interested in this project?

KLI is interested in this project because they are an all-remote team that cannot benefit from in-person upskilling opportunities. The project aimed to design an AI-driven solution for continuous upskilling that could benefit their own employees

What are they hoping to find?

As a research based company, KLI was really looking for the research. While a solution was important, they were more interested in solution evaluations and how people feel about different solutions. With this, they were hoping to find a clear path for designing an AI-driven upskilling solution that addresses the needs of early-career professionals and students.

the Terminology

Upskilling

The process of improving or learning new skills, which may pertain to the individual professionally or personally

Students

People who are juniors or seniors in college who are starting to enter the job market

Early Career Professionals

People at the beginning of their careers, with 1-2 years of work experience

Significant findings

Human Interaction over AI-only Learning

AI is a tool, not a teacher; human guidance is essential. Learners across groups rely heavily on human feedback, peer support, and guidance for motivation and effectiveness.

Trust Gap in AI Capabilities

AI is primarily trusted for technical and repetitive tasks (e.g., summarizing, coding help), but users are cautious about relying on it for deep learning, critical thinking, or nuanced soft skills. Trust is low, especially when users don't already know the topic.

Lack of Practical Application

Students and professionals struggle to translate theoretical knowledge from courses into practical, real-world application. Hiring managers prioritize what applicants can actually do over formal certificates.

Unstructured Learning Creates Drop-off

Major pain points include a lack of structured pathways and unclear learning roadmaps, leading to drop-off in motivation and retention. Early career professionals are overwhelmed by multiple internal tools.

Soft Skills are Undeveloped

Both user groups emphasized the difficulty of developing soft skills (e.g., communication, interpersonal growth) through current tools. Students specifically lacked confidence and job readiness in this area.

AI Personalization Falls Short

Many platforms advertise AI-powered recommendations, but users often find them too generic or misaligned with their actual needs, highlighting a gap between AI promise and user satisfaction.

what this points to?

The problems that we uncovered through research pointed to some solutions, digital and otherwise, that could be implemented to solve the issues.

Hands-On Learning Environment

Implement AI-supported job simulators and real-world task modules so users can "learn by doing" and immediately apply theoretical knowledge to practical projects.

Integrated Soft Skills Development

Incorporate tools such as AI Mock Interviewer role-play simulations and scenario-based learning to organically build soft skills and directly address the pervasive lack of communication and interpersonal skills in new professionals.

Transparent AI and Human Blending

Design systems that explicitly blend AI output with human mentoring or peer check-ins to build user trust by providing transparent, explainable results and clarifying the AI's supportive role.

Our presentation
Key Lime Presentation by Mustafa Arshad
Outcome

Project Deliverables and Key Findings

The final documentation that we handed over to Key Lime

Our FigJam board where we arranged all of our work

Through the project, we achieved the goal of successfully investigating the critical intersection of AI and professional upskilling, moving beyond theory to propose and test tangible, user-validated solutions.


As the people at Key Lime were mainly looking for research, we delivered an extensive research document covering all of our work, along with mid-fidelity design iterations and findings from testing multiple designs.

Through the project, we achieved the goal of successfully investigating the critical intersection of AI and professional upskilling, moving beyond theory to propose and test tangible, user-validated solutions.


As the people at Key Lime were mainly looking for research, we delivered an extensive research document covering all of our work, along with mid-fidelity design iterations and findings from testing multiple designs.

Learnings and Next Steps

Given the rapid increase in its use, it was really interesting learning about perceptions of AI and how people in the industry and final-year students incorporate it. Understanding how to achieve the balance between user trust, usability, and technological feasibility when designing with this technology helped me see how to incorporate it into my process.

The Full Story is currently under construction

Made with Framer, Rive, and a lot of Jolly Ranchers

© 2025 - Mustafa Arshad

Made with Framer, Rive, and a lot of Jolly Ranchers

© 2025 - Mustafa Arshad

Made with Framer, Rive, and a lot of Jolly Ranchers

© 2025 - Mustafa Arshad