Sure! Here’s the translation into American English:
—
In the manufacturing sector, a new technological proposal aims to maximize the information generated in service reports, which is often underutilized in document storage systems. This advancement, designed for Amazon Web Services (AWS) clients, enables the automation of digitization and the extraction of meaningful data from multiple reports using generative artificial intelligence.
The solution is based on Amazon Nova Pro, which integrates with Amazon Bedrock and its Knowledge Bases. This combination aims to generate action recommendations tailored to the observed status of the equipment, relying on a knowledge set that includes expert recommendations. As this tool is used, the knowledge base expands, contributing to continuous improvement in its effectiveness.
Amazon Bedrock is a fully managed service that provides access to high-performance foundational models developed by leading companies in the field of artificial intelligence through a single API. This service also offers various capabilities to create generative artificial intelligence applications, prioritizing security and privacy.
The Knowledge Base system of Amazon Bedrock allows for the construction of Retrieval-Augmented Generation (RAG) workflows that integrate contextual information from the company’s databases. This enables the storage of recommendations from previous reports, improving the accuracy of the foundational model responses.
Traditionally, service and maintenance processes have relied on the manual submission of reports by engineers, a time-consuming method that can lead to operational delays. The new solution promises to optimize this process, giving maintenance teams the ability to:
– Automatically ingest inspection and maintenance reports, enhancing visibility into equipment statuses and pending actions.
– Generate solid recommendations based on engineers’ experience.
– Expand the initial knowledge base with valid and applicable generated recommendations.
– Accelerate maintenance times, thereby preventing unplanned downtime through a centralized AI-powered tool.
To facilitate its implementation, a GitHub repository has been set up that contains deployable code and infrastructure-as-code templates, allowing users to customize the solution within their AWS environment.
This approach encompasses key workflows, including automated ingestion of service reports with Amazon Textract, intelligent generation of recommendations through RAG, and expert validation with Amazon SageMaker Ground Truth. These processes ensure continuous improvement in model performance, thus guaranteeing more reliable recommendations.
With this comprehensive strategy, a transformation in maintenance operations is anticipated, helping to mitigate risks and avoid equipment downtime, highlighting the importance of adopting advanced technologies to optimize operations and create more efficient work environments.
via: MiMub in Spanish
