Workflow · Production

Halliburton reduces seismic workflow creation time by over 95% using Amazon Bedrock and generative AI

The problem

Halliburton's Seismic Engine required manual configuration of approximately 100 specialized tools to build workflows, a process that was time-consuming, error-prone, and required deep expertise, potentially limiting accessibility and efficiency.

Workflow diagram · grounded in source
1
User submits natural language query
trigger
“The backbone of our system is a FastAPI application deployed on AWS App Runner, which handles user queries through a streaming interface.”
2
Intent classification by Nova Lite
routing
“The Intent Router classifies the intent label of the given query by calling Amazon Nova Lite via the Amazon Bedrock API. The large language model (LLM) is given a prompt to produce one of three intent labels: "Workflow_Generation", "QnA"…”
3
RAG-based Q&A answering
ai_action
“For Q&A requests, the system uses Amazon Bedrock Knowledge Bases with Amazon OpenSearch Serverless to provide relevant answers from indexed documentation”
4
YAML workflow generation by agent
ai_action
“the agent selects appropriate tools, determines their correct execution order, and generates a YAML workflow that addresses the user's requirements”
5
Chat history stored in DynamoDB
integration
“To maintain context and enable multi-turn conversations, we integrated Amazon DynamoDB for chat history and interaction logging”
6
Streaming response delivered
output
“The system supports streaming responses for both Q&A and workflow generation, providing immediate feedback to users as the system processes their requests”
Reported outcome

The AI-powered assistant achieved workflow generation success rates of 84–97%, surpassing both new and experienced users, and reduced workflow creation time by over 95% compared to manual processes, while making advanced geophysical tools more accessible.

Reported metrics
Workflow creation time reductionover 95%
Workflow accelerationup to 95%
AI solution workflow generation success rate84-97%
User productivity enhancementover 95%
Show all 9 reported metrics
workflow creation time reductionover 95%
workflow accelerationup to 95%
AI solution workflow generation success rate84-97%
user productivity enhancementover 95%
workflow-building task time reductionby an order of magnitude
Claude Haiku 3.5 simple workflow success rate84%
Claude Haiku 3.5 medium workflow success rate90%
Claude Sonnet 3.5 V2 simple workflow success rate86%
Claude Sonnet 3.5 V2 medium workflow success rate97%
Reported stack
Amazon BedrockAmazon Bedrock Knowledge BasesAmazon NovaAmazon Nova LiteAmazon DynamoDBAmazon OpenSearch ServerlessAmazon Titan Text Embeddings V2Claude 3.5 Sonnet V2Claude 3.5 HaikuFastAPIAWS App RunnerLangChainS3Seismic Engine
Source
https://aws.amazon.com/blogs/machine-learning/halliburton-enhances-seismic-workflow-creation-with-amazon-bedrock-and-generative-ai/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The AI-powered assistant achieved workflow generation success rates of 84–97%, surpassing both new and experienced users, and reduced workflow creation time by over 95% compared to manual processes, while making advan…

What tools did this team use?

Amazon Bedrock, Amazon Bedrock Knowledge Bases, Amazon Nova, Amazon Nova Lite, Amazon DynamoDB, Amazon OpenSearch Serverless, Amazon Titan Text Embeddings V2, Claude 3.5 Sonnet V2, Claude 3.5 Haiku, FastAPI.

What results were reported?

Workflow creation time reduction: over 95%; Workflow acceleration: up to 95%; AI solution workflow generation success rate: 84-97%; User productivity enhancement: over 95% (source-reported, not independently verified).

How is this workflow AI workflow structured?

User submits natural language query → Intent classification by Nova Lite → RAG-based Q&A answering → YAML workflow generation by agent → Chat history stored in DynamoDB → Streaming response delivered.