Originally conceived as Ricardo Lens, this feature was envisioned to streamline the selling journey for our users. Little did we know that this idea would evolve into a top priority for the first half of the year, expanding to improve both buying and selling experiences on Ricardo.
What is Ricardo AI?
Ricardo AI is the polished and advanced version of Ricardo Lens. A feature that allows you to scan an object with the Ricardo app and instantly see listings of the same or similar items currently for sale. Similarly, when selling items, Ricardo AI will generate a product title and description based on the picture you take.
The original concept focused mainly on alleviating the significant hurdles associated with selling items online. Research and user testing indicated that this would be a practical enhancement for users. However, as development progressed, the concept was broadened to include the buying journey, creating a unified and seamless experience for both sellers and buyers.
Key Differences Between Ricardo AI and Ricardo Lens
- Focus on Selling: Initially, Ricardo Lens aimed solely at improving the selling process. The idea was to reduce the time and effort needed to list items online.
- Inclusion of Buying Journey: After the hackathon, the concept was expanded to include the buying journey, allowing users to explore similar items visually and navigate through an endless inventory, making the experience engaging and potentially addictive.
- Innovative Branding: Ricardo AI’s branding diverged from the initial camera-based icon of Ricardo Lens. Through user testing, we discovered that the spark icon resonated with both buyers and sellers, unlike the camera icon which was associated primarily with selling.
The Collaborative Development Process
The journey from Ricardo Lens to Ricardo AI was a collaborative effort between multiple teams, including the original hackathon team, the GM Exploration team, the GM Selling team, and the GM Data team.
- Knowledge Transfer: Julia Fröschner, Product Manager at General Marketplaces, organized a kickoff meeting in January 2024, where the hackathon team shared their insights and experiences with the exploration and selling teams.
- Continuous Feedback: Ash Sharifi, Product Designer at General Marketplaces, acted as an advisor, providing direct feedback to Sergi from the exploration team, ensuring the concept evolved smoothly.
- User Testing: Ash and Sergi Gomez Sanchis, Product Designer, conducted the first user interviews, facilitating a smooth transition and integration of user feedback into the development process.
Technological Backbone of Ricardo AI
One of the initial challenges was whether to launch Ricardo AI as a standalone app or integrate it into the main Ricardo app. The team decided against a separate app to avoid additional marketing costs and user inconvenience. Instead, they opted for a hybrid approach, releasing Ricardo AI to users in the beta track, allowing for rapid iteration without disrupting the regular app release cycles.
Ricardo AI leverages cutting-edge AI technologies, primarily using Google APIs to enhance the user experience. For the buying flow, the API generates image embeddings integrated into the search engine. For the selling flow, the API generates listing information, ensuring accurate and efficient item identification and categorization.
Unique Features of Ricardo AI
- Effortless Selling: Users can sell items as easily as taking a picture, streamlining the listing process.
- Integrated Buying and Selling: The combination of buying and selling journeys through a single interface sets Ricardo AI apart from other AI tools on the market.
- Advanced Search Capabilities: Users can find articles similar to an image, title, or attributes, enhancing the search experience.
User Feedback and Future Evolution
Since its beta launch, Ricardo AI has received overwhelmingly positive feedback. Users rated the experience with an average score of 6/7, placing it in the top 10% of new features for both efficiency and enjoyment. We also anticipate a significant increase in the adoption rate with the upcoming launch of „Selling with Ricardo AI“ in the selling journey.
The selling journey will be available soon, and users will be able to launch Ricardo AI from any product card. Future updates may include features like dictation or sharing web pages with the Ricardo app, reflecting the evolving nature of AI and user habits.
Lessons Learned
- Experiential Convincing: Sometimes, the best ideas need to be experienced to be believed. Hackathons provide this crucial opportunity.
- Mature AI Technology: Experimenting with AI is worthwhile; the quality of results can surpass expectations.
- Cross-Team Collaboration: Hackathons can align different teams on a shared opportunity, fostering innovation.
- Scaled Hackathons: Smaller, team-specific hackathons can be effective in creating momentum and convincing management.
All This to Say
Ricardo AI stands as a testament to the power of collaboration, innovation, and user-centred design. From a hackathon concept to a fully-fledged feature, it has the potential to revolutionise how users buy and sell items on our platform. We look forward to its continued evolution and the exciting possibilities it will bring to our marketplace.