KnowDroids.ai, a developer of AI-driven autonomous agents, was looking for a way to use AI to improve how businesses handle IT workflows. As a Czech company focused on automating tasks like IT support and administration, their vision for smarter, more efficient solutions was something we immediately connected with at Revolgy.
Their goal was to explore how Google Cloud Platform (GCP) and generative AI could enhance their tools while expanding beyond a single cloud provider. Together, we worked on building an AI-powered IT support agent, testing GCP’s capabilities, and exploring different technical approaches along the way. This project wasn’t just about solving a single challenge — it was about helping KnowDroids.ai grow their AI expertise and get ready to support multi-cloud environments.
Making IT support smarter and more efficient with generative AI
At the heart of the project was the development of an AI-powered IT support agent designed to transform how businesses handle IT support tickets. Using advanced AI tools and generative AI techniques, the agent was built to simplify workflows, boost efficiency, and improve customer satisfaction.
“When KnowDroids.ai first approached us, we were excited to hear about their idea of using AI to automate business tasks. They’re creating tools that don’t just solve problems — they make workflows smarter and more efficient. From the start, it felt like a great fit, both for the project and in how well we connected with their team.”
— Teodora Delovska, Project Manager, Revolgy
However, the project was about more than just building an IT support agent — it was an important step in KnowDroids.ai’s journey to grow their AI capabilities and adopt multi-cloud flexibility. With Google Cloud Platform (GCP), they explored tools for data storage, AI model development, and techniques like Retrieval Augmented Generation (RAG). This gave them the chance to see how GCP could improve their solutions and gain hands-on experience with its ecosystem.
The IT support agent served as a test case, automating ticket classification and generating responses for common issues. It demonstrated how AI could make IT workflows more efficient and improve customer satisfaction. The project also gave KnowDroids.ai a chance to compare open-source tools with managed GCP services, helping them decide on the best approach for future AI projects.
The AI tools bringing the IT support agent to life
The project used several advanced AI tools to build the IT support agent, each playing an important role. One of the key technologies was Gemini 1.5, a model from Google used to create human-like responses. For example, when the agent needed to answer a support ticket, Gemini summarized information from a knowledge base into a clear and helpful reply.
“We helped KnowDroids.ai build an IT support agent that uses AI to sort tickets and create quick responses. We combined tools like Langchain to connect everything, Gemini 1.5 to generate replies, and vector databases to store and find information fast. This setup is both easy to use and super efficient.”
— Md Monir Hossain, Google Cloud AI Engineer, Revolgy
Another important tool was Langchain, an open-source framework that connected all the different parts of the system. Think of Langchain as the “glue” that helped the AI components (like the ticket classification model and the database) work together smoothly.
The project also utilized Google Cloud tools like GCS (Google Cloud Storage) for storing data and PostgreSQL with the pgvector extension for managing vector embeddings in the managed GCP setup. Vertex AI played a central role, offering services like Workbench for development, Model Garden for accessing pre-trained models, and tools for tuning, model registry, and pipelines to streamline workflows and improve performance.
The project relied on vector databases to store and retrieve information. These databases are designed to handle large amounts of text data in a way that allows the AI to search and find relevant pieces quickly.
The project tested two setups: open-source and managed GCP setup. Open-source uses a tool called Chroma DB for the database and a fine-tuned open-source AI model for ticket classification and managed GCP setup relies on Google’s Cloud PostgreSQL with pgvector for the database and a fine-tuned Gemini model from Google’s Model Garden for classification. By testing both setups, the team compared the flexibility of open-source tools with the simplicity and scalability of Google’s managed services. This allowed them to see which approach worked best for future projects.
Overall, these technologies made the IT support agent smarter and more efficient. It could quickly classify support tickets and generate accurate responses, significantly reducing the workload for human support teams. The project also demonstrated how flexible and powerful AI tools can be when combined, helping KnowDroids.ai learn what works best for their specific needs.
“When building our AI autonomous agents to serve knowledge-intensive business purposes, we naturally got to the point when we needed to operate multi-model and multi-cloud — and the collaboration with Revolgy accelerated our conquer of new cloud territory and technology, making our ITTicketsDroid a multi-cloud agent; more of his agentic fellows will follow”
— Vladislav Severa, CEO of KnowDroids.ai.
Building real-life solutions with AI
This project showed how modern AI tools can make IT workflows smarter and more efficient. By combining custom and pre-trained models, KnowDroids.ai built a system that’s flexible, scalable, and easy to use. Tools like Langchain streamlined how the AI components interacted with the data, while vector databases made sure information was stored and retrieved quickly and accurately.
The collaboration between KnowDroids.ai and Revolgy went beyond solving a single challenge. It helped KnowDroids.ai strengthen their AI capabilities, explore new approaches, and prepare to support clients across different cloud platforms. This project is a great example of how AI and cloud technology can come together to create real, practical solutions for businesses.
Is your company tackling cloud-related projects? Feel free to reach out for a free technical consultation. We’d be happy to share how our expertise can help turn ideas into successful solutions.
Read also: The executive’s guide to generative AI
