About the client
QUBTRIX is a Polish brand that creates innovative wooden construction blocks compatible with popular systems such as LEGO®. The products combine education, ecology, and aesthetics, supporting the creative development of children and adults. The company focuses on local production, high quality, and functionality, offering solutions that inspire creative play and experimentation.
QUBTRIX's business is based on multi-channel sales – products are available both in online stores (its own e-commerce and marketplace platforms), at trade fairs, and stationary at the Forum Designu in Krakow. The company's dynamic growth and diverse customer base generate numerous questions regarding products, compatibility, orders, and availability.
Challenge
The QUBTRIX team was looking for a way to automate answers to frequently asked customer questions, without the need to involve the support department. The challenge was to create an effective system that would answer questions based on existing documentation and company data. It was crucial to verify the effectiveness of the Retrieval-Augmented Generation (RAG) approach and to leverage the AWS infrastructure in a way that allows for rapid deployment, low maintenance costs, and the possibility of further development of the solution.
Main project goals
- Automating responses to customer inquiries using existing company materials – without the need to involve the team in every interaction.
- Verifying the effectiveness of the RAG (Retrieval-Augmented Generation) approach in the context of chatbots based on large language models.
- Leveraging AWS services to build a scalable and flexible architecture at the lowest possible maintenance costs.
- Facilitating knowledge base updates by automating the process of feeding it with new documents, without the need for manual actions.
- Gathering data from user interactions to further improve the chatbot's performance and better understand their needs.
Situation before the solution was implemented:
Before implementing the chatbot, Qubtrix faced several challenges that affected the efficiency of customer service and the team's work:
- Employees were overloaded with handling repetitive inquiries, which limited their availability to perform key tasks.
- There was a lack of an automated system for updating information, which posed a risk of providing outdated data.
- Customer service was mainly performed manually, which limited the speed and availability of support, especially outside of business hours.
- It was necessary to implement a tool enabling 24/7 and effective communication with customers and to relieve the team.
Our solution
To meet Qubtrix's challenges, we implemented an advanced chatbot based on Amazon Lex technology, integrated with AWS services, which ensures the stability, security, and scalability of the solution.
The chatbot has been equipped with an automatic knowledge base update mechanism through integration with Amazon S3, which allows for continuously enriching and refreshing the information available to users without the need for manual intervention.
Additionally, we introduced the saving of all interactions in AWS CloudWatch, which enables accurate analysis of user behavior and faster detection and correction of potential errors.
An important improvement was also the elimination of the “cold start” problem – the chatbot responds immediately, even after a longer period of inactivity, ensuring a smooth and comfortable experience for users.
Thanks to this solution, Qubtrix can efficiently handle customer inquiries around the clock, relieve its team, and quickly adapt to changing business needs.
Tools and services used
In implementing the chatbot for Qubtrix, we utilized modern cloud solutions and AWS technologies that ensure reliability, scalability, and efficient system operation:
- Amazon Lex – as the main platform for creating and managing the chatbot's conversational interactions,
- Amazon S3 – used for storing documents and files that serve as a source of knowledge for the chatbot and for automatically updating the knowledge base,
- AWS Lambda – enables scalable and serverless processing of queries and handling of the chatbot's business logic,
- AWS CloudWatch – for monitoring and logging user interactions with the chatbot, which allows for analysis and optimization of the system's performance,
- Amazon Bedrock – supports the language model used to generate responses to user queries,
- OpenSearch Serverless – creates and manages a dynamic knowledge base, allowing the chatbot to quickly find the necessary information.
By combining these services, we were able to create an advanced, secure, and easy-to-maintain system that meets the needs ofQubtrix.
Benefits for the client
The chatbot implementation has brought the following benefits to the company:
Reduced time to access information – employees find the data they need faster, which increases their productivity.
24/7 customer support availability – the chatbot answers queries 24/7, improving customer satisfaction and engagement.
Automatic knowledge base updates – ensure that the chatbot always uses the latest information, without the need for manual intervention.
Full control and analysis of interactions – thanks to conversation logs, it is possible to monitor quality and continuously improve the chatbot's performance.
Low maintenance costs – optimized architecture ensures economical implementation without sacrificing quality.
Summary
The implementation of theAI chatbotinQubtrixis an example of a modern approach to automating service and improving business processes. Thanks to the use of AWS technologies and well-thought-out integration, the client has gained a tool that not only increases the efficiency of the team's work, but also increases the level of customer satisfaction thanks to 24-hour support. Optimizing costs and automating knowledge base updates make the solution scalable and ready for future challenges. It is an investment that translates into real savings of time and resources, and also allowsQubtrixto develop dynamically on the market.