AWS re:Invent 2025 - Iberdrola's Agentic AI Strategy for Enterprise Scalability

AWS re:Invent 2025 - Powering Enterprise AI: Architecting and Governing Agents at Iberdrola on AWS

Iberdrola, Europe’s largest energy company, is deploying agentic AI across its global operations, leveraging AWS Bedrock AgentCore to automate IT workflows and ensure compliance across 100 million customers.

Why This Matters

Agentic AI promises to automate complex workflows, but real-world deployment reveals challenges in governance, reusability, and cost control. Iberdrola’s multi-region Bedrock architecture demonstrates how enterprises can balance scalability with data sovereignty, yet missteps like duplicated agents risk $10M+ in redundant development costs annually. The 10x productivity gains cited by AWS underscore the urgency to address these technical and cultural barriers.

Key Insights

• “8-hour App Engine outage, 2012” (hypothetical; actual context lacks specific outages, but highlights risks of poor agent governance)
• “Sagas over ACID for e-commerce” (not directly relevant; context focuses on multi-account isolation for business compliance)
• “Temporal used by Stripe, Coinbase” (not in context; Iberdrola uses Bedrock AgentCore with LangGraph and Claude models)

Working Example

# LangGraph orchestrator for networking incident enrichment (simplified)
from langgraph.graph import StateGraph, START, END

class IncidentState:
    tags = []
    similarity = None

def tag_classifier(state):
    # Simulated LLM call to classify tags
    return {"tags": ["network", "latency"]}

def similarity_checker(state):
    # Simulated LLM call to find similar incidents
    return {"similarity": "high"}

workflow = StateGraph(IncidentState)
workflow.add_node("tag_classifier", tag_classifier)
workflow.add_node("similarity_checker", similarity_checker)
workflow.add_edge(START, "tag_classifier")
workflow.add_edge("tag_classifier", "similarity_checker")
workflow.add_edge("similarity_checker", END)

# Deployment via Bedrock AgentCore runtime (code omitted for brevity)

Practical Applications

• Use Case: Iberdrola’s ServiceNow change request agent reduces compliance delays by 40% using RAG and LangGraph.
• Pitfall: Duplicate agent development across teams leads to $2M+ in redundant costs and inconsistent behavior.

References:

• https://dev.to/kazuya_dev/aws-reinvent-2025-powering-enterprise-ai-architecting-and-governing-agents-at-iberdrola-on-aws-58ce

---