What challenges did Iberdrola face before AI implementation?

Iberdrola needed to streamline multi-phase change management and speed up incident handling. The company wanted to reduce bottlenecks and improve data quality and consistency using contextual intelligence. It also aimed to simplify and accelerate selecting and completing change model requests.

What was the impact of the AI solution?

Reduced processing delays across change and incident workflows. Improved data handling consistency and quality across departments. Enabled faster ticket resolution. Reduced infrastructure complexity and engineering overhead by moving point automations to reusable production-grade agents.

Source

Evidence: Low25/100

Iberdrola enhances IT operations using Amazon Bedrock AgentCore

Use case typeAutomotive operations automationUpdated Jun 13, 2026

Iberdrola, a global utility company, applied agentic AI to IT operations to improve change request validation, incident enrichment, and change model selection. The solution uses Amazon Bedrock AgentCore with LangGraph-based agents, MCP-style tool gateways, identity, memory, observability, Amazon ECR, Amazon S3, Amazon RDS with pgvector, Amazon Bedrock models, Amazon Bedrock Guardrails, ServiceNow integration, and Amazon EKS-based tracing. The implementation runs production-ready agents inside Iberdrola’s enterprise security and VPC environment to support multi-step workflows across departments.

Organization: Iberdrola
Industry: Energy & Utilities
Location: Spain
Published: February 2026

Reported outcomes

Strategic outcomes

Speed & agilityStreamlined change management workflowsCustomer experience & trustImproved data quality and consistencySpeed & agilityEnabled faster ticket resolutionCost efficiencyReduced infrastructure complexity

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Iberdrola
Provider: AWS
Maturity: Unknown
Linked source: AWS Machine Learning Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 4

1IT operations automation
2Incident management
3Change management

Iberdrola needed to streamline multi-phase change management and speed up incident handling.
The company wanted to reduce bottlenecks and improve data quality and consistency using contextual intelligence.
It also aimed to simplify and accelerate selecting and completing change model requests.

Implemented production agent workflows on Amazon Bedrock AgentCore.
Built a sequential change-management workflow with specialized agents for rule extraction, content validation, model analysis, and phase transition.
Built an incident-enrichment workflow with a master agent that routes to specialized agents for tagging, similarity, associating changes, and retrieving context.
Added a conversational assistant for change model selection that returns clickable recommendations and pre-filled forms.
Used Amazon S3, Amazon RDS with pgvector semantic search, ServiceNow integration, Amazon Bedrock Guardrails, and observability tooling to support the workflows.

Technologies

Amazon Bedrock AgentCore Amazon Bedrock Amazon Bedrock Guardrails Amazon ECR Amazon S3 Amazon RDS pgvector ServiceNow Amazon EKS LangGraph MCP

Reduced processing delays across change and incident workflows.
Improved data handling consistency and quality across departments.
Enabled faster ticket resolution.
Reduced infrastructure complexity and engineering overhead by moving point automations to reusable production-grade agents.

Architecture

The architecture uses a layered agentic design with AgentCore Runtime for containerized LangGraph agents, AgentCore Gateway for MCP-style tool access, AgentCore Identity for authentication, AgentCore Memory for session state, and AgentCore Observability for logs and metrics. ServiceNow provides request input, ETL pipelines move operational data into Amazon S3 and Amazon RDS with pgvector for semantic lookup, and Amazon EKS is used for tracing and monitoring. Agents run in a VPC-isolated production environment with VPC endpoints, and Amazon Bedrock models plus Guardrails provide model inference and safety controls.

Sources & evidence1

Evidence: Low25/100Evidence strength