AI-Powered ChatBot
Enhancing Customer Journeys:
Real-Time AI Assistance
ON AG aimed to enhance their customers’ shopping journey by integrating AI to deliver personalized, real-time assistance.
Our mission was to explore the capabilities of AI, develop an MVP, and seamlessly integrate the solution into their existing systems.
Timeline
From explorations to development and launch in the US in 3 Months.
Background
ON AG (ON Running), a leading Swiss sportswear brand renowned for its innovative athletic footwear and apparel, sought to elevate its customer engagement by integrating an AI-driven chatbot into its digital ecosystem. The goal was to harness the capabilities of artificial intelligence and large language models to provide personalized, real-time assistance, thereby enriching the shopping journey for their customers.
In this project, I collaborated closely with the AI and backend teams, leading the initiative from the mobile development perspective. My focus was to ensure seamless integration of the chatbot within the mobile application, aligning with ON Running's commitment to delivering a superior user experience.
This category details the step-by-step approach taken during the project, including research, planning, design, development, testing, and optimization phases.
Research & Planning
We conducted in-depth market research to identify common challenges in e-commerce chatbots, such as poor user experience, lack of personalization, and limited integration capabilities. Our analysis of industry leaders revealed that successful implementations prioritize clear objectives, seamless integration with existing systems, and continuous performance monitoring. These insights guided our approach to developing an AI-driven chatbot tailored to enhance user engagement and satisfaction.
Design & Prototyping
Collaborating closely with ON Running's in-house design team, we prioritized rapid MVP development by focusing on essential features that deliver immediate value to users. Our design approach embraced minimalism and clarity, ensuring an intuitive user experience while maintaining flexibility for future enhancements. By adopting agile methodologies and iterative design principles, we created a scalable foundation that allows for seamless integration of new features as they are validated
Development & Implementation
I developed a streamlined MVP of the AI-powered chatbot, incorporating essential features to deliver immediate value to users. To validate user interactions and feature effectiveness, we developed & implemented an elaborate analytics concept, enabling us to monitor real-time engagement and make data-driven decisions. Additionally, we implemented feature flags to facilitate controlled rollouts, allowing us to test new functionalities with a subset of users before a full-scale launch.
Deployment, Testing & Optimization
The pilot was rolled out to 10% of the US user base over the course of one month. I collaborated closely with the analytics team to analyze user behavior and gather valuable insights. The pilot proved to be a success, and based on our findings, we developed a comprehensive action plan for further development. After this phase, the project was smoothly handed over to an internal team for ongoing maintenance and expansion.
The resulting AI-powered Chatbot Pilot offered a glips into the future of E-Commerce.
Seemless User Communication
Leveraging the power of large language models (LLMs), we delivered a highly personalized chat experience. The chatbot interacted seamlessly with users, creating the impression of a natural, real-person conversation.
Pre-Purchase
The chatbot recommended products dynamically based on the ongoing conversation. Users could provide feedback on these suggestions, allowing the chatbot to learn and refine future recommendations for a truly personalized experience.
Post-Purcahse
The chatbot also provided users with assistance in order tracking and customer support, ensuring a seamless and efficient experience.
Overall, the project was a success, and the business was pleased with the results—so much so that they formed a new internal AI initiative team. We successfully launched the pilot within three months, gathering valuable user feedback and identifying the strengths and weaknesses of AI-based chatbots.
Technologies
E-Commerce, Flutter, Dart, Firebase, Sentry, Google Analytics, Strapi, REST, LLM, Chatbot, Automated Testing with Maestro, Docker, Contentful, iOS, Android