Tatneft Turns 75-Year Knowledge Archive Into AI Assistant “Guru”
The AI system analyzes hundreds of thousands of documents and processes thousands of engineering queries every day.
Russian oil company Tatneft has transformed its 75-year archive of technical knowledge into a digital assistant called Guru, the company’s press service said.
The tool is built on generative artificial intelligence and combines several technologies, including large language models and vector databases.
Guru quickly retrieves information from hundreds of thousands of documents, generates possible solutions by analyzing accumulated expertise, and reduces the workload on specialists by handling routine queries. The digital assistant covers more than 380 knowledge domains and processes thousands of engineering requests every day.
Useful Even for School Students
An additional Guru component described as a smart “librarian” searches through a collection of about 100,000 books, identifies the relevant volumes, highlights key passages, and then feeds them to the language model, which reads the material and produces a solution.
The company added that the corporate neural network also helps school students. Using Guru, they analyze plant health and learn the basics of robotics.
