About the customer
Strawberry, a prominent hospitality company in the Nordic region, operates over 234 hotels and employs more than 20,000 people. Renowned for its focus on quality and innovation, Strawberry is dedicated to exceptional guest experiences. The company upholds its core values of energy, courage, and enthusiasm while embracing modern technology to enhance its services. By fostering a vibrant culture, Strawberry sets new standards in the hospitality industry, ensuring memorable stays for guests across the region.
The Challenge and Goal
Strawberry faced challenges when upgrading its legacy systems to more flexible and lightweight solutions. The shift introduced new information and procedures, making it difficult for employees to find and understand what they needed. Traditional knowledge-sharing methods and internal communication struggled to keep up with these changes. To address this, Strawberry aimed to create a centralised platform where employees could easily access information about the company’s culture, processes, and FAQs. This platform would streamline operations and improve the overall employee experience.
The Solution
Devoteam customises and expands Strawberry’s AI chatbot, Scout, using Google’s Gemini technology. Scout analyses Strawberry’s internal information sources, including documents, websites, and knowledge bases. It integrates seamlessly with common tools like Google Chat and Slack, fitting into Strawberry’s current workflow. Strawberry needed an accessible solution that learns from existing information and adapts to their needs. Scout, built with Google Cloud’s Gemini, understands and responds in English, Swedish, and Norwegian. LangChain helps Scout remember past conversations and adjust to Strawberry’s specific needs. Additionally, function calling enables Scout to retrieve real-time information from Strawberry’s internal systems.
Specific tools used include:
Google Cloud: The chatbot was developed with Google Cloud’s Gemini large language model, allowing it to understand and respond in English, Swedish, and Norwegian while learning from Strawberry’s information sources.
LangChain: This tool creates a context-aware chatbot by considering the history of conversations and Strawberry’s specific environment.
Function Calling: This feature lets Scout access Strawberry’s internal tools and databases directly, providing real-time information and enhancing its functionality.
The Devoteam G Cloud team worked closely with Strawberry to tailor the solution. They focused on integrating Scout with Strawberry’s diverse data sources, customising the chatbot’s personality and tone to match the company culture, and ensuring data security and GDPR compliance. The team followed a collaborative and iterative approach: first, they conducted a discovery phase to understand business requirements and existing systems; then, they developed a roadmap for integration, customisation, and scalability; followed by pilot testing with feedback on accuracy and user experience; and continuous improvement based on that feedback. The rapid deployment of a minimal viable product within two weeks demonstrated early success. High user adoption, with employees quickly embracing Scout and consistent accuracy in responses reduced the burden on support channels and saved time. Strawberry did not have to change its existing tools or platforms, as Devoteam ensured Scout seamlessly integrated with Google Workspace, Google Chat, and Slack, familiar interfaces for employees, boosting adoption further.
The Result
Strawberry has transformed from fragmented information sources and inefficient knowledge-sharing methods to a centralised platform that streamlines information retrieval. Scout a Chatbot AI solution built with Google Cloud, has become a trusted resource for employees. It significantly boosts efficiency and enhances the overall employee experience. This project highlights the successful collaboration between Devoteam G Cloud and Strawberry, showing how innovative solutions can drive operational improvements and make the organisation more data-driven.