What challenges did Halliburton Landmark face before AI implementation?

Creating seismic processing workflows in Seismic Engine previously required manual configuration of about 100 specialized tools. The process was time-consuming, error-prone, and required deep expertise, limiting accessibility and productivity.

What was the impact of the AI solution?

Workflow generation success rates reached 84-97%. Workflow creation time was reduced by over 95%, from minutes to seconds. Complete workflows were generated in about 5.9-16.6 seconds.

Rank#916of 3,107

Source

Halliburton Landmark AI assistant for Seismic Engine workflow creation

Use case typePlanning automationUpdated May 8, 2026

Halliburton Landmark built an AI-powered assistant for Seismic Engine to convert natural-language requests into executable seismic workflows and answer questions from documentation. The solution uses Amazon Bedrock Knowledge Bases, Amazon Nova Lite, Amazon OpenSearch Serverless, Amazon Titan Text Embeddings V2, Amazon App Runner, and Amazon DynamoDB in a conversational workflow-generation and Q&A architecture.

Organization: Halliburton Landmark
Industry: Energy & Utilities
Location: United States
Published: May 2026

Reported outcomes

97%

workflow generation success rateAdoption & scale

84%workflow generation success rate−95%workflow creation time reduction

Strategic outcomes

Employee experienceMade advanced geophysical tools more accessibleSpeed & agilityEnabled conversational workflow creation and modificationScale & capacityImproved productivity across subsurface workflows

Catalog median for adoption & scale deployments: +85.5% across 46 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 1 of 1

1Planning automation

Creating seismic processing workflows in Seismic Engine previously required manual configuration of about 100 specialized tools.
The process was time-consuming, error-prone, and required deep expertise, limiting accessibility and productivity.

Halliburton partnered with the AWS Generative AI Innovation Center to build a conversational assistant for Seismic Engine.
An intent router powered by Amazon Nova Lite classifies requests into workflow generation, Q&A, or general questions.
Q&A uses Amazon Bedrock Knowledge Bases over S3 documentation with Amazon OpenSearch Serverless and Amazon Titan Text Embeddings V2.
Workflow generation uses Claude models on Amazon Bedrock to select among 82 tools and generate YAML workflows, with DynamoDB storing chat history and interaction logs.

Technologies

Amazon Bedrock Knowledge Bases Amazon Nova Lite Amazon OpenSearch Serverless Amazon Titan Text Embeddings V2 Amazon App Runner Amazon DynamoDB Anthropic Claude 3 Haiku Anthropic Claude 3.5 Sonnet

Workflow generation success rates reached 84-97%.
Workflow creation time was reduced by over 95%, from minutes to seconds.
Complete workflows were generated in about 5.9-16.6 seconds.

Architecture

A FastAPI application deployed on AWS App Runner handles user queries through a streaming interface. An intent router powered by Amazon Nova Lite classifies requests into workflow generation, Q&A, or general questions. For Q&A, the system uses Amazon Bedrock Knowledge Bases with Amazon OpenSearch Serverless and Amazon Titan Text Embeddings V2 over documentation stored in S3. For workflow generation, a generation agent using Claude on Amazon Bedrock selects from 82 Seismic Engine tools and generates YAML workflows. Amazon DynamoDB stores chat history and interaction logging for multi-turn conversations.

Implementation partners1

AWS Generative AI Innovation Center

Sources & evidence1

Groundedness: 5/5Type: Blog PostPublished: May 8, 2026Publisher: AWSEvidence: VendorConfidence: Medium

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence