.png)
Introduction
Alfa Laval faced a growing challenge: their experts were spending too much time manually searching and analyzing technical documents, which slowed down operations and risked delaying critical decisions. They needed a solution that would allow teams across the order process to access precise technical, regulatory, and contractual information quickly and reliably.
Partnering with Lettria, they aimed to create a smart document assistant capable of handling complex, multilingual documents while providing accurate answers in natural language. This collaboration combined Alfa Laval's deep industry expertise with Lettria’s advanced GraphRAG methodology to transform document intelligence across projects.
Partnering to Reduce Expert Workload
Their challenge was that searching for and analyzing technical information consumed a significant amount of time from their highly solicited experts. They wanted a solution that could save time for experts and allow all users involved in orders to quickly access key information from large volumes of documents.
Choosing Lettria to Support Their Teams
Lettria’s open approach and genuine interest in their use case convinced them to launch a multi-month proof of concept. The promising results and potential for further improvement led them to continue the RAG journey with Lettria, aiming to industrialize document integration, improve answer quality, and leverage Lettria’s ongoing innovations.
Enabling Quick, Accurate Answers Across Departments
Their objectives included relieving experts from repetitive document review tasks and enabling all departments involved in orders to get fast, accurate answers to specific technical questions using natural language.
How Lettria Supports Their Knowledge Management
Lettria helps them integrate complex documents, model their ontology based on domain knowledge, and optimize prompts to deliver precise, verifiable answers. Here is a sample document of the type of complexity Lettria handled :
.jpg)
A Stepwise Collaboration
Building Adoption Across Teams
In the first year after the POC, they planned to onboard around ten projects to ensure adoption across the organization.
Pilot Testing and Gradual Deployment
The solution was tested on a pilot project involving sales, quality control, engineering, and project management teams. Deployment then continued with additional projects.
Continuous Improvement Through Feedback
User feedback on answer quality is shared with Lettria to diagnose inaccuracies and either improve the platform or guide users on best practices for optimal results.
Measurable Impact of the Partnership
- Reduced time spent per collaborator on reviewing technical data.
- Decreased manual effort in managing files.
- Higher percentage of correct answers.
- Ability to index large, complex document sets per project.
