A company that produces offshore drilling equipment with several offices and over six thousand employees worldwide has innovation and technology at its forefront. Since its founding in the 1950s, the public company has continuously grown their fleet of drilling units, specializing in technically demanding sectors of the global offshore drilling business, while consistently maintaining their records of versatility, safety, and environmental quality. As the company has grown over the past seven decades so has their custody of physical documentation, ranging from human resources records to technical specification documents. Multiply that by the number of offices located around the world, the number of employees, and the number of physical assets and the resulting paper trail is exponential in size. A football field full of paper!
The cognitive and digitization services incorporated in Knowledge Mining technology allow for companies to deeply understand their information, explore it, and uncover insights that were not previously as readily available prior to Artificial Intelligence (AI). Having contracts, memos, presentations, human resource files, schematic documentation, and any other type of hard or soft copy data available at one’s fingertips is invaluable to any company in this fast-paced, digital world.
This major producer of offshore oil rigs found itself hoarding millions of pages of hard copy documents. After 70 years of being in business in an era that has transformed greatly in computing and networking, the company still relied on outdated methods of sharing information within the company itself. In order to access a work order or a rig specification, a request had to be made with a data custodian or archivist who then had to physically go search for those paper files in a warehouse full of filing cabinets, possibly located in a different international location. Lest we not forget to mention the complexity that the sophistication (or lack thereof) of one’s filing system would add to finding a specific file or record. Once a file was located, the document would in most cases be faxed to the requestor. Finding files and sharing them was an inefficient and time-consuming process, extremely burdensome and dependent on man-hours.
It was time for a change and digitizing over 8 million files seemed like the obvious solution. But also utilizing AI, how could the solution encompass a new wave of technology allowing for users across the globe to access and gain insight from data that previously only existed in one office somewhere in the world?
BlueGranite partnered with the company and formulated an approach to create an enterprise search solution by applying knowledge mining to business documents like contracts, memos, schematics, images, et al. Hard copy documents would be targeted first. Imagine scanning and storing millions upon millions of physical documents of varying sizes from different locales! A third-party vendor was contracted to scan documents consolidated at two separate office locations and store them on a secure (SFTP) server accessible to BlueGranite.
We first utilized Azure Cognitive Search to index and OCR (optical character recognition) files as they were being loaded onto the server. The OCR process included indexing words within documents or images, not just indexing the files themselves. Also, as files were landing on the SFTP server, the document management group beta tested and validated the files. BlueGranite built a custom application for end users to access and search through archived records using a number of search and filter criteria. This custom application is available online 24 hours a day from virtually any office location around the world. Data is now at the fingertips of the users who require this information in seconds not weeks or months.
BlueGranite also employed Microsoft Power BI to run reports monitoring the status of the files being processed, to manage errors in processing, and to read storage tables in order to alleviate issues before they arose.
Our client looks forward to implementing the next steps in Knowledge Mining including Form Recognition and Document Classification which would give users the ability to run searches on certain types of files, finding specific pieces of data that are relevant to an issue at hand, for example finding an exact part exists in 12 different schematics. Furthermore, an analytics tool, such as Microsoft Power BI, could be utilized to extract business insights and trends from the now readily available data and expose the results to users at all levels of the company.