Halliburton and the AWS Generative AI Innovation Center have developed a proof — of-concept assistant that converts natural — language queries into executable seismic processing workflows and provides question — answering for Seismic Engine documentation. The conversational interface maps user requests directly to processing steps inside Halliburton Landmark’s DS365 Seismic Engine, potentially speeding interpretation and lowering the barrier to advanced geophysical processing.
The prototype stitches together Amazon Bedrock, Amazon Bedrock Knowledge Bases, Amazon Nova and Amazon DynamoDB with Halliburton Landmark’s DS365 Seismic Engine. A FastAPI application deployed on AWS App Runner handles streaming user queries and translates conversational inputs into configuration settings for the Seismic Engine, removing the need for manual configuration of many pipeline components. Traditionally, configuring seismic processing workflows required time-consuming, expert manual work across roughly 100 specialized tools, which limited efficiency and accessibility. By offering a cloud — native, conversational approach, the project aims to let a broader range of geoscientists and data scientists run advanced processing without deep, tool-by-tool expertise.
Evaluation of the proof — of-concept reported workflow acceleration of up to 95% and a reduction in workflow — building tasks by an order of magnitude. Phillip Norlund of Halliburton Landmark said the collaboration with AWS was “instrumental in accelerating subsurface interpretation workflows.” The team presents this pattern as a blueprint for other organizations looking to streamline complex technical workflows with generative AI.
Sources
Replies (0)
No replies in this topic yet.
