About the Customer
O’Neal Industries (ONI) stands as the largest family-owned network of metals service centers in the United States. They are a leading provider of essential metal products, including steel beams, plates, specialty alloys, components, and tubing, serving a wide range of industries. ONI has a strong commitment to leveraging technology to optimize its operations and enhance its services. This dedication to innovation enables them to make informed, data-driven decisions, ensuring efficient and reliable service within the competitive metals industry.
The Challenge
ONI’s analysis team faced a significant challenge in managing the vast amount of technical documentation they handle daily. This documentation, which includes a mix of PDF, DOC, and XLSX files containing both text and images, plays a crucial role in their analysis and decision-making processes. However, manually extracting and summarizing key information from these files was a highly time-consuming and labor-intensive task. This manual process was not only inefficient but also prone to human error, potentially impacting the accuracy of their analyses and subsequent decisions.
ONI needed a solution that could automate this document processing workflow to improve efficiency and accuracy. Their ideal solution had to meet several key requirements:
- Automated Extraction: Automatically extract text and image data from PDF, DOC, and XLSX files.
- Accurate Summarization: Provide concise and accurate summaries of the extracted information.
- Diverse File Handling: Handle various file types and formats seamlessly.
- Scalability: Efficiently process large volumes of documents to keep pace with their growing needs.
- Timely Insights: Enable faster access to information for timely reporting and decision-making.
The Solution
Onix delivered a generative AI Proof of Concept (PoC) to illustrate how generative AI could automate the extraction, summarization, and analysis of information from ONI’s technical documents. More specifically, our team created an architecture that would allow ONI to extract images from complex PDF documents, generate descriptions of these images, recombine PDFs and images into concise summaries, and turn this output into impactful Google Slide presentations (with the ability to convert these slides into PowerPoint format).
The architecture:
How it works:
- Document Storage: ONI’s documents are securely stored in Google Cloud Storage, providing a centralized and scalable repository for their data.
- Automated Processing: Cloud Functions, a serverless compute service, acts as the orchestrator of the workflow. When new documents are uploaded to Cloud Storage, Cloud Functions automatically triggers the processing pipeline.
- AI-powered Extraction: The core of the solution lies in Vertex AI, Google Cloud’s unified machine learning platform. Vertex AI’s powerful document AI capabilities are used to extract text and images from the PDF, DOC, and XLSX files.
- Intelligent Summarization: Vertex AI goes beyond simple extraction to provide intelligent summaries of the key information within the documents. This helps ONI quickly grasp the essential content without having to manually review lengthy files.
- Centralized Storage: The extracted data and summaries are then stored in BigQuery, Google Cloud’s highly scalable and cost-effective data warehouse. This provides ONI with a centralized and easily accessible repository for all their processed information.
- Reporting and Analysis: BigQuery’s powerful analytical capabilities allow ONI to perform comprehensive analysis on the extracted data. The solution also includes options to convert the data into various presentation formats, further streamlining ONI’s reporting workflows.
Results
The PoC successfully illustrated that by automating the extraction and summarization of information from PDF, DOC, and XLSX files, ONI could drastically reduce manual effort and minimize the risk of human error. Increased accuracy in data analysis and reporting would also be a welcomed byproduct, as well as substantial time savings for ONI’s analysis team. Automation would allow them to focus on higher-value tasks like in-depth analysis and strategic decision-making, ultimately improving overall productivity. Furthermore, this Generative AI solution would enable real-time access to processed data, empowering ONI to generate reports and gain critical insights faster than ever before. This lends itself to greater agility and responsiveness in making informed decisions.
Conclusion
This project showcased the power of AI and Google Cloud in transforming document-intensive workflows. By partnering with Onix and leveraging Google Cloud’s Vertex AI, Cloud Functions, and BigQuery, ONI now has a plan for automating its document processing pipeline and achieving new levels of efficiency, accuracy, and productivity. As ONI continues to grow and evolve, this AI-powered solution will play a crucial role in enabling them to effectively manage their information assets and maintain their leadership position in the metals industry.
