ConverseCart: Redefining Retail with Intelligent Shopping Assistance
Conversations for Conversions
OVERVIEW
Background
ConverseCart is a startup, offering an AI copilot for online retailers, aimed at boosting conversions by enhancing the shopping experience for consumers — making it more enjoyable, trustworthy, simpler, and smarter.
Working with developers and founders, I led the design of the consumer-facing shopping assistant. The focus was on elevating customer satisfaction and driving conversions, employing user research and maintaining a balance between user and business needs.
Team
My Role
2 Developers
1 Founder
1 Researcher
Research
Product design
Dev handoff
4 months
Duration
Client
CHALLENGE
How might we leverage AI to improve the shopping experience for shoppers, making it easier and more enjoyable, while also increasing sales for sellers?
To gain insights into the diverse consumer base of mid-sized retailers, we conducted interviews with 15 shoppers. Additionally, we supplemented our findings with thorough secondary research, leading us to identify and encapsulate three distinct shopping behaviors in the form of personas.
USER RESEARCH
The Shoppers
Current Journey Map
Collaborating with the founders, I identified both potential and existing competitors in the market, spanning across direct, indirect, and varying scales from small to large enterprises. Subsequently, I conducted extensive product testing, scanned reviews, and performed a comprehensive analysis to discern distinctive propositions for ConverseCart.
MARKET RESEARCH
Competitive Analysis
Following thorough user and market research, I distilled the gathered insights into key findings, identified opportunities derived from these findings, and formulated specific recommendations and features to integrate based on the acquired knowledge.
SYNTHESIS
Identifying Opportunities
Why?
Findings from secondary & primary research
Need-Based Shopping is Laborious
Online shopping is too time-consuming and requires effort. People research more and struggle more while searching for niche and expensive products online.
01
Lack of Trust
Small retail shops lack trust and require reviews to verify their products. People can't judge the quality or benefits of one product over another.
02
Strugles wth Search
People struggle with keywords. They expect search engines to understand their queries. They find irrelevant search results frustrating.
03
Frustrations with Chatbots
Limited chatbot capabilities and repetitive pop-ups can be frustrating for users, causing inconvenience and dissatisfaction. Exception: Customer support.
04
What?
Opportunities identified from findings
Provide ‘hidden’ discount codes that are relevant to where the viewer is.
Educate and entice users. ‘Trends?’ ‘Frequently bought?’ ‘Why is this trending?’
Strategic upselling and cross-selling
Urge to spend more time on the website.
Do not hinder the current shopping experience. Easily closed or denied.
Customizations possible on sellers’ end. Understand different types of sellers as well.
Make their job easier: Offer ‘research’ features based on reviews and other comparators
Gain trust: Tell ‘why’ a product is recommended.
Limit open ended questions. Provide cues and options to help them initiate the search.
Give users options to ask specific questions in a structural manner. FAQs? Selections?
Relevant filters and easy sorting.
Let the users navigate the website easily. Integration between CC and seller's website.
Customizable? Increase and decrease the size, see more info, option to see PDP to read more.
Collect people’s contact (email) for seller’s records.
Collect feedback quickly and conveniently
1/10 rating? thumbs up or down?
Do not compete with onsite search functions for entire catalog browsing.
How?
Specific features we could consider
AI to understand synonyms, and what people might mean. Then transparently display search results.
Contextual copy/prompt displayed according to the user’s activity.
‘Find matching sets’ if they add a top the cart.
Keep sentences and paragraphs short to make content appear less intimidating.
Suggest differences or comparisons between different items. Have a 'Compare' button.
A bar to select no of options shown. Less, targeted options (5?). More, browse related products also (~20)
Target abandoning visitors - slightly expand and show discount coupon etc.
Never have a blank chat window. Always guide the user with FAQs, Popular categories etc.
Show editable filters while the user is interacting with chatbot.
Through user testing, internal discussions, and feedback from developers, we explored opportunities and ideas, creating and refining final designs through continuous iterations.
TESTING & ITERATIONS
Solutions
Contextual copy/prompt displayed according to the user’s activity.‘Find matching sets’ if they add a top the cart.
Never have a blank chat window. Always guide the user with FAQs, Popular categories etc.
Show editable filters while the user is interacting with the chatbot.
Suggest differences or comparisons between different items. Have a 'Compare' button.
01. Context helps
Users were more inclined to tap on the assistant with contextual prompts as compared to a standard one.
Iterations →
02. Need for Flexibility
Showing the quantity of content depends on sellers' data, thus only category name will be displayed if images are missing.
Iterations →
03. Prioritizing
Comparion table is too technically complex. For now, comparisons are GPT generated answers with the table view reserved for the next iteration of the MVP
Iterations →
The client, Uncle Reco, is currently piloting the proposed designs and undergoing testing. The launch of an MVP has successfully garnered ConverseCart its initial clients. Similar initiatives are underway for diverse markets and potential clients. I am actively working with ConverseCart, continuously iterating based on user and team feedback.