BACKGROUND

Collaborating with Honda 99P Labs, our team explored how Large Language Models (LLMs) could enhance Honda’s value propositions. We focused on improving interior wayfinding experiences and optimizing attendee interactions at large events. This led us to design an app, “Koo,” which highlights AI-embedded features to significantly enhance attendee experiences while aligning with Honda’s business objectives.

THIS PROJECT DEMONSTRATES:

As the product designer on the team, I focused on creating Koo to augment attendee experiences, striking a balance between user needs and AI strengths. Key contributions included:
‍
🔵 Designing initial wireframes for the MVP and conducting early user testing. I refined the design to a mid-high fidelity prototype, incorporating user feedback to ensure the functionalities reflect user needs.
‍
🔵 User Interface design with a focus on meaningful tasks: I emphasized AI-enabled tasks that provide genuine value from the user’s perspective, ensuring that each feature aligned with practical attendee needs.
‍
🔵 Presentation to Honda 99P Labs: Demonstrating the user flow and final design, we received positive validation, confirming that Koo aligns with Honda’s goals and enhances the attendee experience through AI-driven features.

CHALLENGE

This project challenged our team to build a new product from scratch (0-1), requiring a foundation of extensive AI research and deep user and stakeholder understanding that includes:
‍
🔵 Researching AI capabilities to identify which tools could truly enhance the product and add unique value.
‍
🔵 Aligning with stakeholders to clarify expectations and ensure the product would meet broader business goals.
‍
🔵 Gaining insights into user needs and pain points to tailor the solution to a genuine problem.

The core challenge has been identifying a real problem where AI can deliver unique value, moving beyond surface-level applications to uncover meaningful ways AI could transform the user experience. Balancing innovative AI use with practical value to users has been a continuous journey, refining the product vision to ensure it’s both technically feasible and resonant with user needs.

final design OVERVIEW (mid-fidelity)

In Koo, we introduced four core AI-powered features designed to transform the tradeshow experience through automation and augmentation, creating a seamless, personalized journey for attendees. These include:

Personalized Itineraries

Leveraging onboarding questions, Koo uses algorithms to generate customized itineraries tailored to each attendee’s preferences and goals. These schedules are dynamic, adapting in real time based on attendee feedback and live event changes, ensuring each attendee has a schedule that suits their evolving needs.

Dynamic Indoor Navigation

Using Large Visual Model, Koo offers indoor navigation capabilities to help attendees navigate complex event venues. The feature ensures they arrive at sessions on time without the stress of finding their way, significantly enhancing their experience and saving valuable time.

Automated Information Management

To allow attendees to fully engage with the event, Koo automates note-taking and information management. It captures and organizes insights from sessions, generating summaries that attendees can review later, thus allowing them to stay focused on networking and learning.

Enhanced Networking

Koo’s AI algorithms facilitate networking by identifying potential contacts based on shared interests and business relevance. By highlighting meaningful connections, Koo enriches the attendee experience, helping them forge valuable relationships that might not have happened otherwise.

SCOPE

Understanding and Aligning with Honda 99P Labs’ Vision:

At the project kickoff, our team initially received an unclear project scope from the Honda 99P Labs team. Through focused discussions and alignment meetings, we clarified the objective:

Explore an AI-powered Attendee Assistant solution designed specifically for large trade shows and events.

‍

research

🔵 Understanding AI Capabilities and Limitations

To build an effective solution, we first needed a grounded understanding of machine learning (ML) concepts and current AI applications. Through in-depth research, we clarified what AI can and cannot achieve, addressing common misconceptions and establishing best practices to build trust in AI-powered features. This foundation helped us to make informed decisions about realistic applications for the Attendee Assistant.

🔵  Aligning Technology with Attendee Experience

To gain firsthand insights, our team conducted field research at the Pittsburgh Auto Show, using the AEIOU (Activities, Environments, Interactions, Objects, Users) framework. We also interviewed over 50 attendees to learn about their motivations, experiences, and pain points. From these interactions, we identified three critical areas for our MVP:

While one auto show doesn’t capture every aspect of visitor experience, it provided a valuable starting point. From this direct exposure, we synthesized key findings and identified crucial factors in the trade show experience to investigate further.

🔵  Transforming Research Insights into the MVP Concept

Trade shows are complex, high-energy environments that often test the limits of attendees’ organizational skills. Navigating large spaces, accessing information, and managing time effectively are challenges that can define or derail the trade show experience.
‍
‍To inform our design process, we initiated an artifact analysis to benchmark existing products in the market. One standout competitor was Cvent, which offers a comprehensive event management app that aligns closely with our vision for Koo. By examining its features and user experience, we gained valuable insights into potential strengths and weaknesses in the current market landscape.

Koo is designed to streamline and elevate the attendee experience by automating traditionally difficult or tedious tasks and enhancing opportunities for engagement and personalization. Automating Complex or Unpleasant Tasks includes first-draft itinerary generation, notes synthesizing, seamless navigation Augmenting tasks for enhanced experiences includes meaningful network, knowledge acquiring, personalized recommendations.

Through a combination of automation and augmentation, Koo transforms traditionally cumbersome tasks into seamless, engaging experiences, allowing attendees to focus on connecting, learning, and enjoying the event.

Product Benchmarking & Wireframe

In addition to Cvent, we explored various other applications that incorporate specific features relevant to our design goals including MazeMap, Google Lens, This exploration helped guide our design decisions, ensuring that we addressed user needs effectively while differentiating Koo from existing solutions.

MVP user testing with experts

After developing low-fidelity wireframes, we conducted user testing with industry experts to evaluate the key features of Koo. During these sessions, we guided participants through the product and solicited their feedback on functionality, usability, and overall design.The user testing highlighted several important insights, which we incorporated into the next stages of our design process.

‍Familiar UI Design
Based on user feedback, we adjusted the itinerary interface to resemble apps like Google Calendar, making the design more intuitive and familiar for users.

Onboarding with progress bar to learn user preferences

Users preferred a questionnaire over a chatbot for onboarding. We implemented pre-selected default answers to streamline the process, improving onboarding speed and effectively gathering user preferences.

Simplified Map Feature
adding details that support navigation
Users found a 3D map unnecessary, so we replaced it with a simpler 2D version similar to Google Maps, focusing on usability without excess complexity.

Add networking feature with Linkedin integration

Since many users use LinkedIn for networking, we introduced a barcode-sharing feature that aligns with user preferences while enhancing the platform’s networking capabilities.

Enhanced Note-Taking Experience

To improve the note-taking feature, we added the ability to create to capture and transcribe images to text, and organize notes into files and folders.

Final design in Mid fidelity

Given the time constraints of the project, I focused on developing the app to a fidelity level that effectively showcased the main features and interactions, while emphasizing the integration of AI. This approach ensured that our presentation to the client clearly conveyed the product's core functionalities and the innovative ways in which AI enhances the user experience.

Reflection

Koo project was a transformative experience that deepened my understanding of user-centered design and the integration of AI in enhancing user experiences. One of the key takeaways was the importance of thoroughly understanding user needs through direct engagement, such as our field research at the auto show. This approach allowed us to ground our design decisions in real-world insights, ensuring that the features we developed were not only innovative but also practical and relevant.

Additionally, the iterative nature of our design process highlighted the value of user testing. Receiving feedback from industry experts not only validated our assumptions but also illuminated areas for improvement that we might not have identified otherwise. This collaborative approach fostered a sense of ownership and commitment to delivering a product that truly enhances the attendee experience.

About designing AI product
AI is fantastic for tasks like recommending personalized content or recognizing images, but sometimes simple rules or manual input can do the job just fine without complicating things. Heuristics or manual control can often create better experience. In this case, it's important to find the intersection of user needs & AI strengths in order to solve a real problem in a ways that AI addes values.