Chatso's Coupon Feature

Enabled agents to generate coupons from product recommendations, driving more effective upselling.

Product designer / Breadstack Tech / 2023

The Overview

About Chatso

Chatso is a live chat platform developed by Breadstack Tech for e-commerce businesses, allowing agents to engage with customers in real-time while they shop online. Beyond facilitating conversations, Chatso also acts as a sales-driving tool, equipping agents with features that guide purchasing decisions and surface upselling opportunities.

Project Summary

As part of Chatso's sales-driving capabilities, the platform offers features like product recommendations and coupon generation. Initially, these features operated independently, so agents who wanted to upsell a recommended product with a coupon had to create the coupon separately, adding friction when speed and momentum were crucial.

To address this gap, we expanded the coupon feature to allow agents to generate coupons directly from recommended products. The goal was to reduce context switching, preserve momentum, and make upselling more seamless during live interactions.

How it Started

Identifying the Problem

Agents occasionally recommend products in Chatso as part of guiding customers toward a purchase. When customers hesitated, offering a coupon was a common next step to encourage conversion.

However, since coupon generation existed outside the recommendation workflow, agents had to switch contexts and manually configure a coupon, disrupting the flow of conversation during time-sensitive sales interactions.

This highlighted an opportunity to reduce friction by enabling agents to create coupons directly from product recommendations, helping them act faster and keep momentum in the chat.

Defining the Project Scope

Key Constraints

This project was positioned as a feature expansion, building on existing systems and components rather than creating something entirely new.

With the high-level strategy and solution direction already defined by our Design Lead, my role was focused on translating that direction into a clear interaction and interface that fit within the current product experience.

Because much of the groundwork was already in place, I worked on a tight timeline while navigating the following key constraints:

The solution needed to live within the existing Product Recommendation component
The existing design system and visual language needed to be maintained
Coupon logic had to align with the current coupon generation system

Establishing the Approach

Before moving into detailed design, I aligned with stakeholders on guiding principles to ensure the solution was intuitive, consistent, and workflow-friendly. The work focused on three core areas:

How to make the coupon action immediately clear

How to integrate the action seamlessly

How to optimize the end-to-end flow

Designing & Iterating

Starting with the Flow

To minimize risk and maintain familiarity, the interaction flow closely followed the existing Send Coupon/Coupon Generator experience, with small adjustments to make it more contextual and efficient:

Reviewing this flow with the team confirmed it was clear, low-risk, and easy to implement. Reusing the existing coupon experience reduced learning overhead while improving speed during live conversations.

Visual Exploration

With the flow defined, the main design challenge was surfacing the coupon action within the Product Recommendation list without adding visual clutter or distracting from the live chat.

I explored three interaction patterns:

Inline Action: Hover-based approach with inline actions
Dropdown: A dropdown that revealed additional actions per product
Card Interaction: An expandable card interaction using elevation and shadow

Each option was evaluated against clarity, speed, and alignment with the existing design system.

Final Refinements

After reviewing the concepts with the team, the dropdown approach emerged as the strongest solution. It clearly communicated that each product had additional actions while keeping the interface compact and easy to scan.

The pattern also aligned with existing behaviors elsewhere in the product and avoided drawing unnecessary attention away from the live chat experience.

The Solution

Launching the Feature

Once the interaction pattern was finalized, I refined UI details, documented edge cases, and handed off final designs to development.The result was a unified workflow that enables agents to generate coupons directly from recommended products.

By reducing context switching and embedding incentives into the recommendation flow, the experience became more intuitive and optimized for upselling opportunities

Quick Coupon Generation: Agents can access a coupon action directly from any recommended product during a chat, allowing them to create a product-specific incentive in seconds.

Streamlined Discount Configuration: Selecting the action opens the coupon configuration panel with the product preselected, enabling agents to quickly define discount values and other details.

Seamless Sharing with Customers: Once configured, the coupon and product link are sent together in chat. Customers receive both the product page and the discount code, ready for checkout.

The Impact

From Conversation to Conversion

After launching the expanded coupon feature, clients saw strong and measurable results:

25% chat-to-coupon rate
On average, 1 in 4 customer conversations resulted in a product-recommended coupon being generated, signaling that the feature integrated naturally into agent workflows.

23% coupon redemption rate
Nearly a quarter of sent coupons were redeemed, directly contributing to conversion.

$118 average order value
Orders with a product-recommended coupon averaged $118, nearly 3x higher than the $44 average for orders without coupons, supporting higher-value transactions.

By embedding coupon creation directly into the recommendation workflow, we removed friction at a key sales moment and improved a single interaction within the existing flow, leading to measurable gains in both conversion and order value.

Thanks for reading! Head back home here →