A company serving the agriculture and construction industries was manually reviewing documents containing information on its 30,000 different parts and products. Infosys' AI-powered solution aimed to streamline this process.

The company had relevant information spread across hundreds of different format documents and vendor websites. They wanted to extract this information securely, as some of it was confidential. The main goal was to design a solution that could automate this extraction process.

Key Challenges

  • Information relevant to the parts and products was spread across hundreds of documents that appeared in several different formats (PDF, Word, Excel, scan, etc.), while also appearing on the vendor's website
  • The company wanted to extract relevant information from those documents and websites in a safe and secure manner (some of the information was confidential). Some of this extraction was underway but was being conducted semi-manually or manually

Ready to experience?

TALK TO EXPERTS
Line

The Solution

AI-powered automated text extraction that provides summaries for thousands of products

Infosys started by carrying out extensive preprocessing on the parts and products information, which was fed into a large language model (LLM) to improve accuracy. Building on this, tools harnessing the power of Infosys Topaz were used to automate text extraction. These tools provide summarized specifications for thousands of products and spares, leveraging Microsoft Azure OpenAI along with an interactive UI hosted on AWS.

Accelerating access to information with an AI-driven solution

  • Preprocessing of information through an LLM
  • Use of Infosys Topaz for automated text extraction
  • Leveraging Microsoft Azure OpenAI and AWS for interactive UI hosting
Line

Benefits

Diminished need for manual review: A 90 percent reduction in manual efforts by automating text extraction and content generation

Diminished need for manual review: A 90 percent reduction in manual efforts by automating text extraction and content generation

Reduced processing time: Processing time plunged from 480 minutes per document of 60+ slides to under one minute

Reduced processing time: Processing time plunged from 480 minutes per document of 60+ slides to under one minute

Low cost: The solution could be reused at no additional cost for any of the company’s summarization needs

Low cost: The solution could be reused at no additional cost for any of the company’s summarization needs