AES, a global energy company, uses Claude on Google Cloud’s Vertex AI to automate safety audits and accelerate their mission of transforming the future of energy. By leveraging Claude’s advanced natural language processing capabilities via Vertex AI, AES has:

• Increased audit accuracy by 10-20%
• Generated audit results 96% faster
• Reduced audit costs by 99%

The challenge of scaling renewable energy asset management

As AES transitioned to renewable energy, their success created a new challenge: how to efficiently manage and maintain a vastly more complex network of power-generating assets. Dr. Sean Otto, Senior Director of Data Science and Analytics at AES, explained the scale of this challenge, saying, “With traditional power plants, you might have one large turbine generating a gigawatt of power. But with wind energy, each turbine only generates 1.5 megawatts, so you need 75 to match that output. Instead of maintaining one piece of equipment, we’re now managing 75 separate assets.”

This exponential growth in assets required a complete reimagining of AES’ safety and maintenance processes. The company conducts about 1,550 internal safety audits annually, typically performed by employees whose primary roles are not auditing. These audits, which could take up to two weeks to complete, were becoming an unsustainable drain on time and resources. AES needed an innovative solution that could dramatically reduce the effort required for audits while maintaining or improving their accuracy.

Why AES chose Claude on Google Cloud’s Vertex AI

For AES, Claude’s capabilities and Vertex AI’s robust infrastructure were key to scaling their safety audit solution securely and efficiently. Security was paramount. Running Claude on Vertex AI provided the enterprise-grade security AES needed for their sensitive data. As Dr. Marwan Sherri, Data Scientist at AES, noted, “Integrating Vertex AI with Claude made it easier to get approval to start API calls. Because it’s already integrated and all security issues are addressed.”

Another critical factor was performance. Sherri highlighted the importance of speed and accuracy, saying, “Claude is accurate and quick. We need this process to be 100 times faster than a human.” Claude’s ability to follow instructions precisely was essential for AES’ multi-agent framework, and its support for higher API call volumes enabled better synchronization of their agent system.

Integration was seamless since AES was already using Vertex AI. As Sherri explained, “We’re very familiar with Vertex AI. To launch anything through Vertex AI, we need an API call integration so we can integrate the tool inside Vertex AI with the project-based configuration to solve related issues like permission levels or the service account.” This familiarity accelerated development and deployment of their AI-powered safety solution.

How Claude on Vertex AI powers AES’s AI auditor

AES developed an AI-powered auditing system using Claude on Vertex AI to automate their internal safety audit process. The system consists of three layers of AI agents:

1. Document processing agents that analyze audit documents and create tasks
2. Task breakdown agents that simplify complex audit tasks
3. Report generation agents that compile results and produce final audit reports

This multi-agent system powered by Claude on Vertex AI can process hundreds of pages of safety documentation, evaluate compliance, and generate comprehensive audit reports.

By using Vertex AI, AES leverages Google Cloud’s infrastructure for seamless deployment and scaling.

Transforming safety and efficiency

AES improved their safety operations by integrating Claude on Vertex AI into their audit processes. Audit reports that took two weeks to complete can now be generated in about an hour. This 99% reduction in audit time allows AES to conduct more frequent and thorough safety checks across their expanding renewable energy assets.

The increased speed hasn’t compromised quality. AES has seen a 10-20% improvement in audit accuracy. Dr. Sherri noted, “These agents can’t make mistakes due to their focus on details.” This accuracy contributes to AES’ core value of safety, ensuring reliable protection for employees and assets.

The cost savings have been equally impressive, with audit expenses reduced by 99%. This reduction allows AES to allocate resources more effectively, potentially enabling greater investment in renewable energy expansion and innovation.

Looking forward to AI’s role in accelerating renewable energy

AES sees generative AI as a transformative force in the energy industry. Otto emphasized, “The power of GenAI is in process improvement.” By using Claude to automate complex processes like safety audits, AES can redirect their focus to innovation and expansion in renewable energy.

Dr. Otto envisions AI driving positive change in a traditionally conservative industry, explaining, “AI is important because it’s driving innovation and pushing people to question traditional ways of doing things. We’re reimagining processes that have been unchanged for decades.” With Claude’s advanced capabilities on Vertex AI, AES aims to streamline operations, manage more renewable assets efficiently, and ultimately accelerate the world’s transition to sustainable energy.