Overview
With the rise of AI agents simplifying everyday tasks, I explored how this shift could improve a product I use often — Zepto, the 10-minute grocery delivery app.
Through this side project, I reimagined Zepto's mobile experience by introducing an AI-powered shopping assistant — one that could proactively help users reorder, explore new items, and complete their grocery list with minimal effort.
Role: Product Designer
Type: Self Initiated
Problem Space:
Consumers shopping for groceries typically exhibit two cognitive states:
Routine behavior (reordering familiar items quickly)
Exploratory behavior (discovering new or complementary products)
Current UX flows on Zepto cater primarily to self-driven routine behaviors through lists, recommendations, and search. However, they place cognitive load on users to navigate categories, recall past needs, and manually build carts — even when their shopping patterns are predictable.
There exists an opportunity to shift from user-initiated actions to agent-assisted, goal-driven shopping, reducing friction and increasing both user satisfaction and conversion rates.
Design Challenge:
How might we design an AI-powered shopping agent within Zepto’s consumer app that proactively assists users in completing their grocery orders, enhances decision-making efficiency, and drives higher daily order rates — while maintaining Zepto’s core principles of speed, simplicity, and trust?
The agent must:
Recognize patterns in user purchase behavior.
Predictively assist users in cart building and completion.
Support natural, conversational interactions for tasks like “Reorder my staples” or “Suggest ingredients for pasta night.”
Checkout and pay
This was just a thought experiment — but with the pace AI is moving, who knows? Maybe it’s not that far off.😁