OTTO Launches New AI Assistant Systems: Built on Gemini, Focusing on Product Discovery and After-Sales Service

[Ebrun Exclusive] April 18th news: Recently, German e-commerce platform OTTO announced the launch of two new artificial intelligence systems, targeting the shopping decision-making and after-sales service processes respectively, marking a further advancement in its "AI-driven shopping experience" initiative.

A core component of this launch is a conversational shopping assistant developed in collaboration with Google. This system is built on Google Gemini technology and focuses on product discovery and personalized recommendations.

Unlike traditional search methods that rely on keyword filtering or category navigation, this assistant allows users to express their needs through natural language, whether via text or voice input, and the system can understand these requests and provide corresponding product suggestions.

Specifically, the assistant enables users to describe complex needs in everyday language, such as making purchase requests based on specific life scenarios or personal conditions. After receiving a request, the system refines the requirements through multi-turn dialogue and matches them with product data for recommendations. This interactive approach is designed to mimic the consultation process with sales staff in physical stores, aiming to increase user confidence in purchase decisions while reducing the likelihood of incorrect purchases and returns.

Technically, the assistant leverages the Gemini model to process natural language input and integrates it with OTTO's own product database. Currently, the system can access data from over 18 million products on the platform, including product descriptions and structured attribute information. Future plans include incorporating product review data to further optimize recommendation accuracy.

Additionally, the assistant can recognize dialects and colloquial expressions, enhancing the naturalness of interactions. According to official disclosures, this feature is currently available in a beta version and will undergo continuous iteration and expanded coverage in the coming months.

Alongside the shopping assistant, OTTO also introduced an AI system for customer service scenarios. Unlike the former, this system was developed in-house and is primarily used to handle order inquiries, delivery status, return processes, and invoice-related issues.

It operates 24/7 and responds to user requests within seconds. For highly standardized queries, the system can process them automatically; for complex situations requiring human judgment, requests are forwarded to customer service agents.

Architecturally, this customer service assistant is directly integrated into OTTO's internal service systems and operates using a combination of various language models and structured knowledge bases. It also features an intelligent distribution mechanism that categorizes user requests and directs them to the most appropriate handling channels, ensuring precise connection with specific support teams when necessary.

With the implementation of these two systems, OTTO is gradually building an AI framework that covers the entire shopping journey. In fact, OTTO's exploration of AI-assisted recommendations has been ongoing for years. As early as 2023, the company began related testing with Google.

OTTO CEO Boris Ewenstein stated that artificial intelligence is fundamentally changing digital shopping, with the long-term goal of creating a unified AI environment spanning product discovery, comparison, decision-making, and after-sales services.

Ebrun will continue to track developments on this topic. To learn more about related information, please scan the QR code to follow the author's WeChat.