Gerdau: uso de IA generativa en la gestión de snapshots de EBS en AWS 

1. Sobre la empresa: Gerdau 

Gerdau is one of the largest Brazilian companies and one of the world’s leading steel producers, with global industrial and commercial operations. Its technology landscape is highly complex, with extensive use of cloud computing, multiple AWS accounts, and distributed teams, which requires:

• Gobernanza sólida y control de costos
• Alta eficiencia operativa a escala
• Prácticas maduras de FinOps
• Visibilidad clara sobre recursos en la nube activos, inactivos e innecesarios

2. Desafíos

Como parte de su estrategia de gestión de costos en AWS, Gerdau identificó un desafío recurrente relacionado con los snapshots de EBS Snapshots..

Although EBS snapshots are incremental, they consume storage in Amazon S3. Over time, old and forgotten snapshots may accumulate data that is no longer needed, resulting in unnecessary monthly costs. Many organizations discover that a significant portion of their AWS bill comes from unused or unmanaged snapshots.

The main challenges were:

• Identifying EBS snapshots not associated with any AMI
• Finding snapshots that had been stored for more than 15 days
• Ensuring that snapshots generated automatically by AWS Backup were excluded
• Lack of a centralized view across multiple AWS accounts
• Need to understand ownership and responsibility, using tags and account-level data

This scenario led to:

• Hidden and recurring cloud costs
• High manual effort for analysis
• Increased operational risk when deciding what could be safely deleted

3. The solution

PierCloud implemented an automated, AI-powered solution using its platform.

3.1 Cloud Compliance Analyzer (CCA)

A custom rule was created in the Cloud Compliance Analyzer with the following criteria:

• Identify EBS snapshots without an associated AMI
• Include only snapshots older than 15 days
• Exclude snapshots created automatically by AWS Backup
• Consolidate data across all AWS accounts
• Include tags and owner information to support governance and accountability

This transformed a complex, manual process into a repeatable and automated FinOps control.

3.2 Lighthouse Intelligent Analyst (LIA)

La Lighthouse Intelligent Analyst (LIA) was used to accelerate rule creation:

• A natural language prompt described all technical requirements

• The AI:

  • Identified the relevant snapshot parameters

  • Built the rule logic

  • Generated the necessary code for execution in the CCA

This significantly reduced development time and enabled fast iteration.

4. Results Achieved 

The solution delivered clear operational and financial benefits for Gerdau.

Reducción de Costos

• Precise identification of snapshots eligible for deletion

• Elimination of recurring, unnecessary storage costs

• Greater control over a commonly overlooked component of AWS billing

Eficiencia Operativa

• Significant reduction in manual analysis

• Clear, reliable list of snapshots that could be safely removed

• Lower risk of human error or accidental deletion

Improved Governance

• Visibility by account, owner, and tags

• Easier coordination with responsible teams

• Cleaner and more organized cloud environments

Practical Use of Artificial Intelligence

• AI accelerated rule creation and refinement

• Fast adaptation when new requirements emerged

• Scalable approach for future FinOps and compliance rules

5. Lessons Learned

Early technical requirement definition is critical: Some requirements, such as deeper tag-based ownership tracking, were identified in later iterations and could have been incorporated from the first version.

Rapid iteration adds value: Even with evolving requirements, AI-enabled development allowed quick and controlled improvements.

Forgotten snapshots represent hidden cloud costs: This case reinforced that unmanaged snapshots can account for a meaningful portion of AWS spending.

Final Message

With PierCloud, Gerdau transformed hidden and recurring snapshot costs into a fully automated, governed, and AI-driven FinOps process—reducing expenses, improving operational efficiency, and strengthening cloud governance.