Industrial Sector

Burckhardt Compression: Chat-based search of specification documents with GenAI

Within 5 weeks, Burckhardt Compression and Zühlke realised a proof of concept for a chatbot based on generative AI and thus identify the potential to increase efficiency and save thousands of hours of manual work.

picture of a manufacturing hall

The project at a glance

Burckhardt Compression faces significant time and resource consumption due to extracting information manually from product specification documents.

With the vision of leveraging Generative AI, Burckhardt Compression aimed to streamline this information retrieval processes, improving operational efficiency and resource utilisation.

Zühlke and Burckhardt Compression successfully developed a proof of concept for a chatbot solution, which identifies the potential of simplifying internal processes and realising substantial time savings.

Manufacturer Burckhardt Compression provides compression services and solutions for various industries. A key challenge facing the company is the task of manually extracting data from a myriad of product specifications submitted by customers. This labour-intensive process consumes a significant amount of time and resources.

Determined to innovate and streamline its operations, Burckhardt Compression enlisted the help of Zühlke, harnessing our expertise in data and AI solutions to pioneer a transformative solution.

The vision: leveraging generative AI to increase operational efficiency

Each year, Burckhardt Compression employees together spend a significant amount of hours manually searching through product specification documents, usually characterised by a variety in structure, content, and references. This process encompassed diverse compression products, making it an arduous and time-consuming task.

With the vision of leveraging Generative AI, Burckhardt Compression aimed to explore the feasibility of a chat-based system, empowering users to effortlessly verify standard requirements or pose customised inquiries. The primary goal was clear: fasten the information retrieval processes, thereby enhancing operational efficiency and resource utilisation.

Technical and business consulting ensures efficiency and impact

Zühlke developed a proof of concept (PoC) for Burckhardt Compression for this generative AI-based system in a standardised framework using Microsoft Azure AI services. The team not only conceptualised the solution but also evaluated the technical feasibility for scaling in a productive environment, ensuring seamless integration into Burckhardt Compression's processes.

Combining technical expertise with insightful consulting, Zühlke collaborated closely with Burckhardt Compression, aligning the chatbot's functionalities with user requirements and business KPIs, such as response quality and response correctness among others. This approach ensured that the chatbot integrates with existing workflows and aimed to maximise the solution's efficiency and impact.

AI-driven innovation boosts efficiency and opens new opportunities

In a remarkable period of only 5 weeks, the Zühlke team delivered a fully operational PoC, deploying it seamlessly within the Microsoft Azure ecosystem. Leveraging the expansive capabilities of Azure OpenAI, the chatbot was augmented with a customer-specific knowledge base, tailored to Burckhardt Compression's documents.

The result is a technical prototype that, once developed into an MVP, has the potential not only to simplify internal processes but also to promise substantial reductions in manual data extraction efforts by approximately 10%.

Burckhardt Compression is now in a position to harness the power of AI-driven innovation, tap into new efficiency potential and lead the business into the future.