Talkn.ai

Talkn is an AI-assisted e-commerce platform that helps users discover products through a conversational interface.

Many users don't know exactly what they want to buy, making traditional search and filtering ineffective. Talkn addresses this by guiding users through product discovery using an AI chat assistant, helping them clarify preferences and find relevant products in a more intuitive and engaging way.

I worked on the UX and UI of this project from August 2023 to November 2023.

My responsibilities included designing user flows, creating the UI, developing the branding, and collaborating closely with developers.

The team consisted of 1 designer (me), 1 frontend developer, 2 backend developers, 1 marketing specialist, 1 sales representative, and the product owner.

Users often struggle to interact with e-commerce platforms when they don't have a clear idea of what they want to buy. Most online shopping experiences rely heavily on search and filtering, which require users to know exactly what they are looking for.

AI-powered chat assistants offer flexibility, but often lack structure and guidance, making product discovery feel unclear and overwhelming.

Design a guided product discovery experience for users without a clear intent.

The goal was to use AI chat to help users explore options, refine preferences, and make decisions, while keeping the interaction clear and easy to follow.

To better understand the problem space, we focused on users who struggle with product discovery when they don't have a clear intent.

Two key personas were identified:

• Emma — unsure what to buy, looking for inspiration, guidance
• Daniel — has a general idea but needs help refining options

While both personas were important, the main focus was on Emma, as she represents the biggest gap in current shopping experiences.

To understand where users struggle, we mapped the journey of discovering a product without a clear intent.

The journey shows key moments where users feel uncertain, overwhelmed, or unable to move forward, especially at the beginning and during decision making.

Pain points were prioritized based on their impact on users and the effort required to address them.

Users don't always know what to type or how to start interacting with AI.

To better understand this problem, I looked at how similar platforms solve it.

• Reviewed AI-assisted shopping experiences (e.g. Shopify, Shopwith AI)
• Analyzed how users interact with chat-based product discovery

Key insights

This helped identify a few key gaps in how users are supported.

• Chat lacks guidance
• AI is reactive
• Product discovery is not structured

These insights defined the direction for a more guided and structured experience.

Based on these insights, I explored different approaches to understand what works best for guiding users through product discovery

We combined the most effective elements into one entry experience to support different user behaviors.

The design focuses on helping users start easily, refine their needs through chat, and explore relevant products in a structured way.

Users can go from "I don't know" to exploring in seconds.

Users felt confused when chat and product results appeared together.

It wasn't clear how chat worked or how it influenced the results, so many users ignored it and continued browsing as usual.

Based on this, I focused on making the interaction more structured and easier to follow. The goal was to make chat feel like a core part of the experience, not something optional.

I explored different ways to combine chat and product results without overwhelming users. Through low-fidelity prototypes and testing with 5 users, it became clear that keeping both visible at the same time works best.

The final design uses a layout with chat on the left and products on the right, combining guidance with a familiar browsing pattern.

Users can follow a clear flow, refine their preferences, and see how their input affects results in real time.

Since the product was in an early stage, these metrics can be used to track user behavior and engagement.

The final solution addressed both key pain points by making it easier to start and creating a more structured interaction between chat and product results.

Users can now begin quickly, stay guided through the experience, and clearly understand how their input shapes the results.

We also explored features for future iterations:

• Saved items (shopping lists)
• Product comparison
• Search history
• "Search similar" functionality

This project changed how I think about designing with AI.

AI is powerful, but without structure it creates confusion. Design is what makes it usable.