Human Computer Interaction (HCI) TEST

Monish Shrestha
UNIVERSITY ID: 0362091
The integration of Artificial Intelligence (AI) and Machine Learning (ML) has transformed the e-
commerce industry, enhancing personalization, customer service, and operational efficiency.
AI/ML technologies such as recommendation systems, chatbots, and dynamic pricing algorithms
have significantly reshaped e-commerce platforms globally. This analysis reviews three key
AI/ML applications in e-commerce: Amazon's recommendation system, Alibaba's AI-driven
chatbots, and eBay's dynamic pricing algorithms. Below is the SWOT analysis based on the
services and functionalities provided by these implementations.

1. SWOT Analysis on Three AI/ML Applications in E-Commerce:

In the evolving landscape of e-commerce, AI and ML technologies have become integral to

improving user experience, streamlining business operations, and boosting sales performance.
This section presents detailed SWOT analysis for three major AI/ML implementations in e-
commerce: Amazon’s AI-Powered Product Recommendations, Alibaba’s AI-Driven Customer
Service Chatbots, and eBay’s Dynamic Pricing Algorithms. These implementations leverage
AI/ML technologies to provide enhanced services and functionalities for both consumers and
businesses in the e-commerce domain.

1.1 Amazon’s AI-Powered Product Recommendations

Amazon’s recommendation engine is one of the most advanced systems in e-commerce, powered
by AI and machine learning techniques such as collaborative filtering, deep learning, and
predictive analytics. This system analyzes data from previous purchases, browsing history, and
user behavior to deliver highly personalized product recommendations. By doing so, it helps
users discover new products that align with their interests, enhancing their shopping experience.
Additionally, the engine supports cross-selling and upselling by suggesting related or
complementary products, which boosts the average order value. The system also adapts in real-
time, continuously adjusting recommendations as the user interacts with the platform, ensuring
that the suggestions remain relevant and timely.
1.2 Alibaba’s AI-Driven Customer Service Chatbots

Alibaba has integrated AI-driven chatbots across its platforms, including [Link] and
AliExpress, to provide seamless customer support. These chatbots leverage natural language
processing (NLP) and machine learning to understand and respond to a wide range of customer
inquiries, such as assisting with order tracking, processing returns, and offering personalized
product recommendations. This AI implementation improves operational efficiency by reducing
the need for human customer support agents while ensuring high service availability. The
chatbots offer 24/7 customer support, helping at any time of day or night. They also offer real-
time order tracking and help manage returns and refunds. Furthermore, the chatbots suggest
personalized product recommendations based on user preferences or previous shopping behavior,
enhancing the overall shopping experience.
1.3 eBay’s Dynamic Pricing Algorithms

ShopSavvy is a mobile and web application that uses AI and Machine Learning to provide real-
time price comparisons and dynamic pricing insights for consumers and retailers. Users can scan
barcodes or search for products to compare prices across online and offline stores. AI algorithms
analyze price data to find the lowest available prices, while ML models predict price drops and
send alerts. For businesses, ShopSavvy offers competitive intelligence, helping retailers track
competitor pricing and adjust their own dynamically. The platform also personalizes shopping by
recommending products based on user preferences and geo-location. With integration across e-
commerce and physical stores, ShopSavvy enables automated, competitive pricing strategies in
real-time.
Strengths Weaknesses Opportunities Threats
Amazon Personalized Privacy concerns Integration with Stricter data
recommendations due to extensive AR & voice privacy
improve user data tracking, assistants, regulations,
experience, Increases Filter bubble effect Expansion into Competition
sales through cross- limits new product social from other AI-
selling & upselling discovery and commerce and driven platforms
and Scalable & data- Requires constant AI-driven trend and Customer
driven. updates to stay forecasting distrust over
relevant. excessive
personalization.
Alibaba 24/7 automated Limited ability to Multilingual & Preference for
support reduces handle complex voice support, human customer
human workload, queries, lack of predictive support,
Cost-efficient for human touch in customer issue competition
businesses, Self- customer resolution and from more
learning AI improves interactions and cross-platform advanced AI
responses. needs frequent integration with assistants and
updates to enhance e-commerce. privacy concerns
accuracy. over storing
conversations.
ShopSavv Real-time price Retailers may Expansion to Legal challenges
y tracking across block access to global e- on dynamic
multiple retailers, price data, commerce pricing fairness,
AI-powered price frequent price markets, AI- competitors like
forecasting helps changes may powered price Google
users find the best confuse customers negotiations for Shopping &
deals and enhances and relies on third- users and Honey.
shopping decisions party data sources. integration with , retailer
with personalized AI shopping resistance to
alerts price
 transparency.

2. Proposed AI/ML Features

Based on the weaknesses and limitations identified in section 1, the following AI/ML-based
features are suggested for enhancing e-commerce AI applications. that can improve e-commerce
platforms by enhancing user experience, optimizing pricing strategies, and making AI-powered
assistance more reliable. The proposed features are as follows:
1. Privacy-Focused AI Recommendations
One of the major concerns with AI-driven recommendation systems is data privacy and the filter
bubble effect, where users are repeatedly shown similar products, limiting discovery. To address
this, an on-device AI recommendation system can be implemented. This feature allows the AI to
analyze user behavior locally on their devices without sending personal data to external servers,
reducing privacy risks. Additionally, to avoid repetitive recommendations, diverse AI models
can be trained to balance personalized suggestions with trending and new products, ensuring
users explore a variety of items rather than just those they’ve previously browsed.
2. Hybrid AI Chatbot with Human Handoff
AI chatbots are efficient for handling large volumes of customer queries, but they often struggle
with complex or emotionally sensitive inquiries, leading to customer frustration. A hybrid
chatbot model can bridge this gap by automatically detecting when a query requires human
intervention. Using sentiment analysis, the chatbot can assess the tone and intent of the
customer's message and seamlessly escalate the conversation to a live customer support agent
when needed. This ensures that basic inquiries (e.g., order status, return policies) are handled
quickly by AI, while complicated issues (e.g., refund disputes, technical problems) receive
personalized human support.
3. AI-Powered Price Transparency & Predictions
Dynamic pricing is a powerful AI tool used by e-commerce platforms, but frequent price
fluctuations can confuse or frustrate customers. To address this, an AI-powered price
transparency feature can be introduced, where users can see why a product’s price has changed
in real time. The system would explain factors like demand surges, stock availability, and
seasonal trends, helping customers understand pricing adjustments. Additionally, AI-powered
price prediction models can be integrated to notify users when prices are expected to drop,
allowing them to make informed purchasing decisions and increasing trust in the platform’s
pricing system.
4. Multilingual & Voice-Enabled Shopping Assistant
A major limitation of current AI chatbots is language barriers, making it difficult for global users
to communicate effectively. To solve this, an AI-powered multilingual shopping assistant can be
developed, allowing customers to search for products, track orders, and compare prices in their
preferred language. Additionally, integrating voice-enabled AI would allow customers to use
natural language commands to perform shopping tasks. For example, users could say, "Find me
the best deal on wireless headphones," and the AI would return a curated list of options based on
real-time pricing and user preferences. This feature would make online shopping more accessible
and user-friendly, especially for those who prefer voice interaction over typing.

5. AI-Driven Smart Negotiation Tool

Price-sensitive customers often abandon their carts when they feel they’re not getting the best
deal. To enhance engagement and increase conversions, an AI-powered negotiation tool can be
introduced. This feature would allow customers to propose a price, and the AI would analyze
demand, inventory levels, and historical pricing data to suggest a real-time counteroffer. This
creates a dynamic shopping experience similar to haggling in physical stores. Retailers can also
set predefined negotiation limits, ensuring profitability while allowing customers to feel they’ve
secured a personalized discount, leading to higher purchase rates and customer satisfaction.

3. Architecture Diagram

Figure 1: Architecture Diagram

3. Proposed Architecture
The proposed architecture for the mobile solution and the technologies needed to implement it.
This architecture supports AI/ML-based features for an e-commerce platform, focusing on
enhancing user experience, optimizing pricing strategies, and integrating AI-powered assistance.
1. User Interface (UI) Layer:
 Technology: Flutter
 Explanation:
The UI Layer is responsible for presenting the application to the user, ensuring a
responsive, interactive, and native-like experience. Flutter is chosen for its cross-
platform development capabilities, which allow a single codebase to work across both
Android and iOS devices. This makes the development process quicker and the user
experience more consistent. Flutter supports rich UI components for building features
like AI-based recommendations, dynamic pricing displays, and chatbot interactions.
2. Mobile API Layer:
 Technology: GraphQL
 Explanation:
GraphQL is selected to allow efficient and flexible data querying. With GraphQL, the
mobile app can request exactly the data it needs, preventing unnecessary data transfer.
This is especially useful for an e-commerce app that needs to provide real-time data, such
as product details, pricing, and order status updates, and integrate AI-powered features
like recommendations and price predictions. GraphQL’s flexibility makes it ideal for
integrating diverse data sources and dynamically updating product information.
3. Backend Services Layer:
 Technology: [Link] (with Express)
 Explanation:
The Backend Services Layer is responsible for processing requests from the mobile app
and managing core functionalities such as user management, product information, order
processing, and more. [Link] with Express is a robust combination, especially for
handling a high number of simultaneous requests. It supports non-blocking, asynchronous
operations, making it perfect for e-commerce platforms that require high performance
and scalability. Express provides a simple and flexible framework for routing and API
management, making it easy to connect with other backend services.
4. AI/ML Engine Layer:
 Technology: TensorFlow (Python)
 Explanation:
The AI/ML Engine Layer is where all machine learning models will be hosted and
executed. TensorFlow is a powerful open-source framework used for developing deep
learning models. In this case, TensorFlow will be responsible for powering features such
as personalized product recommendations, pricing predictions, sentiment analysis for
chatbots, and more. TensorFlow’s ability to handle large datasets and its flexibility in
deploying models for both real-time and batch processing will ensure that the AI-
powered features are accurate and responsive.
5. Database Layer:
 Technology: MongoDB
 Explanation:
The Database Layer will be powered by MongoDB, a NoSQL database known for its
flexibility, scalability, and high performance, particularly when dealing with large
volumes of unstructured data. MongoDB is perfect for storing e-commerce data such as
product catalogs, user profiles, orders, reviews, and AI model-generated data. It is
designed to scale horizontally, which allows it to handle large amounts of transactional
and user-generated data. Its document-based structure supports flexible schema, enabling
quick iterations and adjustments to the data model as the platform evolves.
6. External Data Integration Layer:
 Technology: MongoDB & RESTful APIs
 Explanation:
The External Data Integration Layer is responsible for connecting the application to
external services and data sources. MongoDB can be used here to store and manage
external data that the application integrates with, such as external pricing data, competitor
product listings, and more. It will serve as the data repository for information pulled from
other systems or APIs. In addition, RESTful APIs will be used to fetch external data,
ensuring smooth data flow from third-party services (e.g., payment gateways, product
catalogs from suppliers, or external pricing comparison tools).
7. Security and Authentication Layer:
 Technology: OAuth 2.0 & SSL/TLS
 Explanation:
The Security and Authentication Layer ensures that the application is secure and that
user data is protected. OAuth 2.0 will provide secure authorization, allowing users to
authenticate using trusted third-party services like Google or Facebook. SSL/TLS
 encryption will be implemented to ensure secure communication between the mobile app
and the backend server, protecting user data during transactions and API calls.
8. Cloud Infrastructure Layer:
 Technology: Amazon Web Services (AWS)
 Explanation:
AWS provides a scalable, secure, and reliable cloud infrastructure for hosting the
application. Services like EC2 for compute, S3 for storage, Lambda for serverless
computing, and CloudFront for content delivery will be used. AWS will ensure that the
platform can handle high levels of traffic, scale efficiently as needed, and maintain a high
level of performance globally.