Leverage Knowledge From Over 70,000 Pages of Engineering & Scientific Documents
A case study by AurumNova
Background
An aerospace company develops propulsion simulation software. Over the years, their engineers have curated a digital library of research papers, textbooks, and articles while engineering solutions. The library is stored on the cloud and contains over 70,000 pages of text and images.
The Challenge
Needle In The Haystack
For new projects or maintaining a current project, engineers find it hard to find the correct document, let alone the right page containing the necessary information.
Inconsistent Formats
A good deal of information is found in the images and graphics present in those documents, such as charts, diagrams, and photographs with annotations.
Uncertain Search
Formulating the appropriate question when faced with an engineering issue can be difficult, given the numerous possible solutions.
Our Results
- Every engineer has quick access to a vast knowledge base, which enhances their confidence, accuracy, and productivity.
- The team assured that their documentation has a meaningful impact and dedicates more effort to organizing their documents further enhancing the system.
This drive for positive outcomes is central to AurumNova. Please contact us to discover how AI can significantly enhance your team's capabilities.
How We Did It
Every Page Matters
Given that we want information retrieved to be verifiable, the knowledge ingestion pipeline made sure that every page is indexed and identifiable.
Visual Information
Azure AI Search was leveraged to OCR images found on every page. This means that if there is text in any images, they are indexed in the AI's search database.
Verifiable Answers
When engineers ask questions, their query is enhanced by AI to retrieve relevant documents using Azure AI Search. AI answers their questions with citations to the documents and their pages.