🛍️ Shoppr - Agentic Marketplace

About the project

🤔 What is it?

Shoppr is meant to be a consumer marketplace that leverages AI agents to redefine how people find, evaluate, and purchase products online. In it, users can search for products in natural language, get agent-curated results, find active coupons, and even identify products from pictures, all through a streamlined, agentic-powered experience.

🧠 Rationale

Marketplaces have been a staple of the internet for decades, and yet, the core experience of finding what you want hasn't changed much. You open the app, you type a keyword, you scroll through hundreds of results, you compare tabs, you hunt for coupons elsewhere, and somewhere along the way, you either buy something you didn't need or give up entirely. Where is the real thought on making this process smarter? Where's the meaningful usage of AI beyond slapping a chatbot on top of an existing flow? This is where Shoppr came in, a personal exploration into what a marketplace built around agentic intelligence, from the ground up, should actually look like.

🎯 The challenge

• Design a marketplace experience where AI agents are the core, not a feature bolted on top;

• Find the right visual and interaction language for agentic results that feel trustworthy and transparent;

• Create a purposeful, streamlined experience that respects the user's intent rather than exploiting impulse buying;

• Cover all the core flows of a marketplace, from onboarding to post-purchase, with the agentic layer present throughout;

Developing the concept

🔬 Research and analysis

Since this project sought to understand the current landscape of consumer marketplaces, so we could identify where the standard experience falls short and where agentic features could add the most value, a good starting point was studying the biggest players in the space. Within the wide selection of options, we chose eBay, Mercado Livre, AliExpress, and others, as these represent a broad spectrum of approaches: from the maximalist, stimulus-heavy environments of AliExpress, to the more utilitarian flows of eBay. By checking them out, we got a clear picture of what was working, what was overwhelming, and most importantly, what was missing.

![Research placeholder]

🔍 Finding the edge

After some solid research, the next step was defining the key principles and features that would shape Shoppr's direction. Many concepts were brought up, and here are the ones that stood out:

• Purposeful over impulsive: The biggest insight from our research was understanding the spectrum between maximalist and minimalist marketplace experiences. Platforms like AliExpress are engineered for impulse buying: heavy visuals, endless scrolling, constant promotional noise. This works for them, but it also means the user is always being sold to, rather than helped. Since Shoppr's core is agentic, the whole experience naturally gravitates towards purposeful shopping. Users come with intent, and our tools help them achieve it efficiently. A cleaner, more focused visual environment wasn't just a stylistic choice, it was the only one that made sense;

• Agentic results, human language: One of the main gaps identified across all platforms studied was the disconnect between how people think about what they want, and how they have to express it through filters, keywords, and categories. Shoppr bridges that gap by allowing users to describe what they need in plain language, with the agent handling the interpretation, curation, and even the reasoning behind its picks;

• Transparency as a feature: Agentic tools live and die by user trust. For that reason, every agent result on Shoppr surfaces a "Why this?" affordance, giving users full visibility into the logic behind each recommendation. No black boxes, no guessing, just clear, reasoned output;

• Coupons and deals, handled: One of the most universally painful parts of online shopping is the coupon hunt. Shoppr folds this into the agentic layer entirely, with the agent proactively finding and applying available deals, both at the search stage and at checkout, creating a satisfying payoff that reinforces the platform's value at the exact right moment;

• Visual search: To push the searching experience even further, Shoppr also supports image-based queries, allowing users to upload a picture of a product and have the agent identify and match it to available listings, closing the loop between inspiration and purchase;

Understanding the user

👤 Using personas

Now with a solid grasp obtained by researching and defining concepts, we were ready for the refinement step: creating our persona! As a quick reminder, personas are vital for understanding who your target user is, and what pain points you will be resolving for them. Your ideas and concepts brought up in the previous stage of research will be tested!

👨‍💻 Alex, 29 years (Mindful shopper)

"I've been trying to cut down on unnecessary spending lately, and honestly, most shopping apps make that really hard. Every time I open one, I end up buying three things I didn't plan on getting. I usually know exactly what I want before I open the app. I just want to find it quickly, make sure I'm getting a good deal, and move on with my day, is that too much to ask? I shouldn't need to open five tabs, scroll through hundreds of listings, and manually google coupons just to buy a pair of running shoes…"

💭 Motivations

• Wants to find specific products quickly and without friction;

• Wants to make sure they're always getting the best available deal;

• Values a shopping experience that respects their time and intent;

💔 Pain points

• Gets overwhelmed by the visual noise and impulse-buy engineering of most marketplace apps;

• Wastes time hunting for coupons and comparing listings across multiple tabs;

• Doesn't trust algorithm-driven recommendations that feel more like ads than actual help;

Defining the user flow

📖 User stories

After understanding our persona, we created the user stories, which is the final step to tying our new design proposals with our end-user needs. Some examples below:

• Agentic search: "As a user, I want to just describe what I'm looking for in my own words and have the app find the best options for me, without needing to set up a dozen filters";

• Proactive deals: "I would love for the app to just tell me when there's a coupon available, rather than making me go and find it myself";

• Trustworthy results: "If an agent is picking products for me, I want to understand why it picked them. Just show me the reasoning";

• Visual search: "Sometimes I see something I like and I just want to find it online. I shouldn't have to describe it in words if I already have a picture of it";

• Streamlined checkout: "I want to go from finding something to buying it as quickly as possible, with no surprises at the end";

🫚 User flow

Having all of these done, we created our definitive user flow! This design tool shows a simplified view of how we expect the user to interact with Shoppr's core agentic experience, from first open to post-purchase:

🔗 Shoppr Flow

Working on the visuals

😎 Setting the mood

With our flows mapped, the next step was defining a visual style, our compass through all the UI decisions throughout the rest of the process. For Shoppr, the key words are: clean, easy on the eyes, and purposeful, with a minimalist approach that keeps the focus on the products and the agent output, never on the interface itself;

🎨 Colors

When it comes to colors, the research conclusions made the direction clear. A warm off-white background, paired with high-contrast black type and a restrained green accent, keeps the interface feeling approachable without being loud. The green does double duty: it's the brand color, and it's also the language of value. Discounts, agent highlights, deals, and interactive elements all speak in the same tone. No noise, no competing palettes;

✍️ Fonts

For fonts, the choice was a clean, modern sans-serif that delivers great readability across all screen sizes and contexts, from dense product listings to large onboarding headlines. The type system needed to carry both the functional and the expressive side of the interface without effort;

🫟 Shapes

For shapes, a consistent rounded border radius language was applied throughout, keeping the interface friendly and approachable. Cards, buttons, inputs, and chips all share the same softness, creating a visual coherence that makes even complex screens feel easy to navigate;

The results

Covering all core flows of the platform, here are some of the key experiences delivered:

1️⃣ Agentic search that understands you

Users can describe exactly what they need in plain, natural language. The agent interprets the request, curates the best results, and goes beyond the brief by adding extra value, like suggesting comparable brands the user hadn't considered. Users can also upload a picture of any product and have the agent identify it and surface matching listings. See something you like, find it in seconds;

![Result placeholder]

2️⃣ A clean foundational experience

Even with all the agentic tools offered by Shoppr, users can still find that same familiar core experience of category browsing, seeking for the best prices and the best-selling products. The extra tools are there to add a delightful extra mile to an already complete user journey.

![Result placeholder]

3️⃣ Results with reasoning

Every agent output surfaces a "Why this?" dropdown, giving users full transparency into the logic behind each recommendation. Shoppr doesn't work with random algorithm-based results. Trust isn't assumed, it's earned, one result at a time.

![Result placeholder]

4️⃣ Coupons, handled

From the search stage to the checkout screen, Shoppr's agent proactively finds and applies available deals. The payoff, "Your coupon has been applied!" paired with the discount line in the order summary, is the kind of moment that makes a platform feel like it's genuinely working for the user;

![Result placeholder]

The outcome

Since Shoppr is a personal exploration rather than a live product, traditional outcome metrics aren't the measure here. Instead, this is where we look forward at the OKRs (Objectives and Key Results) and KPIs (Key Performance Indicators) that would guide Shoppr if it were to move beyond concept stage:

• Agentic adoption rate: The primary health metric for this platform would be the percentage of sessions that involve at least one agent query, rather than manual browsing alone. A high adoption rate signals that users trust and return to the agentic flow as their primary way of shopping;

• Query-to-purchase conversion: Tracking how often an agent search leads to a completed purchase, compared to manual browse sessions, would validate the core hypothesis: that purposeful, intent-driven shopping converts better than impulse-based environments;

• Coupon engagement: Measuring how often agent-surfaced coupons are applied at checkout, and the impact on cart abandonment rates, would quantify the value of the proactive deals feature in real terms;

• "Why this?" interaction rate: Monitoring how often users expand the reasoning affordance on agent results would give direct insight into whether transparency is driving trust or being ignored, and inform how much explanation is the right amount;

• Retention vs. competitor benchmarks: Ultimately, a platform built around purposeful shopping should see stronger return visit rates tied to specific needs, rather than passive browsing. Tracking session intent alongside return frequency would be the long-term proof of concept;

You can check a few core screens from this project on the Figma link below:

🔗 Shoppr Preview

If you like this project, or even got some ideas on where to take it next, let me know in our soon-to-be scheduled interview 😉