A Practical Model for Bringing Responsible AI into Real-World Operations
The rapid expansion of generative artificial intelligence (AI) and AI agents is reshaping how organizations operate, make decisions, and deliver services. As companies navigate this transformative landscape, there is a pressing need to transition from theoretical frameworks to practical operational models that demonstrate responsible and scalable use of these technologies.
The specification developed by AENOR in collaboration with Microsoft serves as a robust applied framework, providing verifiable criteria for implementing responsible AI controls in operational environments. This framework emphasizes the importance of:
- Incorporating technical, operational, and organizational controls from the design stage.
- Integrating data quality, technical measures, control mechanisms, and protection throughout the lifecycle of AI systems.
- Generating verifiable evidence that provides objective data to support the responsible use of AI.
Certification Applied to Real-World Use Cases
The certification obtained by the Indra Group validates not only the definition of controls but also their effective application to specific use cases where AI agents developed using Microsoft Copilot Studio technology are utilized.
Two notable systems that have received certification include:
- Onboarding Agent: This system assists new professionals joining the technology group, streamlining their integration into the organization.
- Regulatory Compliance Supervisor Agent: This system verifies and supports compliance with the ISO/IEC 20000 management system, ensuring continuous improvement.
These agents have undergone a rigorous validation process to ensure operational control, traceability, and responsible use, focusing on the technology and the operational lifecycle of the agents.
The Indra Group Case Study: From Experimentation to a Scalable Corporate Model
Carmen Bauset, Head of Systems Governance at the Indra Group, emphasizes that the challenge lies not in the AI model itself but in how it is designed, configured, and operated. The proliferation of agents in environments like Microsoft Copilot Studio highlighted the necessity for a uniform framework across the organization to ensure end-to-end traceability, consistency in controls, and standardized evidence generation.
In response, the Indra Group has developed a proprietary methodology for the design, development, and operation of AI agents. This methodology ensures traceability, control, and the generation of objective evidence throughout the entire lifecycle of AI technologies. It structures the lifecycle into six phases:
- Understanding the business and relevant requirements
- Understanding data and information
- Data preparation, cleansing, and labeling
- Agent design and modeling, including technical configuration
- Internal evaluation and pilot testing with stakeholders
- Iterative deployment, maintenance, and continuous improvement
Proprietary and Aligned Methodology for Managing AI
The AENOR certification recognizes a proprietary methodology integrated within the Indra Group’s Enterprise AI Operating Model (EAOM). This operational framework is designed to manage the industrialized scaling of AI agents in a standardized and measurable manner.
The certification verifies that the controls and mechanisms defined for the design, development, and operation of AI systems are effectively applied in certified real-world use cases. This assessment not only focuses on the technology employed but also evaluates how it is managed and used throughout its lifecycle, ensuring responsible AI usage based on Microsoft’s Responsible AI technology and principles.
Collaboration and Driving the Ecosystem and Innovation
The success of this initiative has been bolstered by close collaboration among Microsoft, AENOR, and the Indra Group. This joint working model has enabled the alignment of technological capabilities with the necessary requirements for control and certification.
Active participation from professionals across the three organizations has fostered a cohesive team environment, underscoring the unique contributions of individuals in the age of AI.
With this certification, the Indra Group positions itself as a leader in the responsible deployment of artificial intelligence. It demonstrates that scaling the use of AI agents can be achieved in an industrialized, controlled manner that aligns with regulatory and market standards. AENOR’s “Responsible AI Technology based on Microsoft tools” certification has emerged as a key solution for organizations aiming to advance AI adoption confidently, transitioning from experimentation to actual operation with tangible evidence and results.
This initiative marks a significant step in the Indra Group’s evolution towards a business model where people and AI agents collaborate seamlessly, transforming processes, scaling capabilities, and accelerating value creation.
