– Case Study –
Automating Database Creation from Research Papers
Leading Education Technology Company Seeks to Automate Reference Database from Research Papers Using a Custom AI Solution by Innodata
A leading education technology company faced the challenge of creating a reference database by indexing terms from geographical research papers, including titles and abstracts. They were seeking an AI solution to assist in identifying relevant index terms based on a provided thesaurus to ensure an accurate representation of the content.
50% Training Time reduced
80-90% Process Automation
To address the challenge, Innodata employed a two-step solution. The first step was to fine-tune the Gen AI model using the provided thesaurus, enhancing its ability to comprehend geographical terms effectively. Next, Innodata developed a streamlined procedure to process geographical content using prompts for precise outputs on index terms from the research paper inputs. Finally, the index terms were converted into the required output file format.
This custom solution resulted in an automated workflow of the customer’s indexing process. Automation of 80-90% of the index term generation significantly reduced the manual effort required, improving overall efficiency and accuracy. With the GPT model’s enhanced capabilities, the throughput of processing files increased by 30%, leading to quicker database creation.
The streamlined approach reduced the time required to train new subject matter experts by an impressive 50%. This efficiency gain allowed for quicker onboarding of new team members, contributing to ongoing increased productivity and cost savings. Ultimately, Innodata’s custom AI solution enabled the customer to create a comprehensive reference database with precise index terms derived from geographical research papers efficiently and accurately.