Ayesa Digital adds GenAI to Iberdrola’s procurement bid evaluation and management

Develops an intelligent platform to optimize the multinational utility’s tendering processes, improving efficiency, agility, and reliability.

Ayesa Digital has developed a generative AI–based platform to optimize tendering processes at Iberdrola. It leverages large language models and machine learning to streamline document categorization, assess risks based on templates, and perform technical and economic pre-evaluations.

The solution also generates evaluation criteria, comparisons, risk analyses, and simulations. It enables intelligent searches across more than 100,000 documents per year and delivers real-time comparative analyses, enhancing efficiency, agility, and reliability.

The project aimed to accelerate the tendering process and make evaluations more efficient by enabling rapid access to key information, centralizing assessments, and promoting data-driven decision-making. It operates across four essential areas:

• Intelligent classification: natural-language searches, automatic summaries, alerts for missing documents, and rapid identification of risks (cybersecurity, subcontracting, etc.).
• Technical criteria: AI generates evaluation criteria based on specifications, proposes improvements, and detects ambiguities.
• Technical evaluation: compares bids against a generated template, highlights discrepancies, and assesses their technical impact.
• Economic evaluation: facilitates cost comparisons, scenario simulations, detection of financial anomalies, and commercial benchmarking.

This reduces time, costs, and manual errors; improves traceability and quality in the procurement process; increases objectivity; and enables analysis of 100% of the documentation.

Applied Innovation and Best Practices

The procurement transformation project stands out for applying generative AI in a tangible way to a critical process. It has now become a digitally assisted workflow, supported by algorithms from criteria definition through final comparison. Semantic search engines and dynamic criteria generation are integrated to suggest improvements, detect ambiguities, and tailor tender documents to real needs.

Through RAG (Retrieval-Augmented Generation), the AI uses up-to-date, contextualized information to compare bids, analyze inconsistencies, and assess technical and economic impacts that previously required many hours of effort. Machine learning enables the tool to be refined iteratively based on historical data, improving reliability. SAFe was adopted for global coordination and Scrumban for daily execution, enabling continuous iteration, incorporation of feedback, and a sustained focus on business value. This reinforces technological innovation and consolidates best practices in digital transformation management.

Use of Technologies

The solution is built on a modern, secure, and scalable ecosystem. Based on AWS, it uses SageMaker to train AI models, Bedrock for generative models, and RDS PostgreSQL for data management—combining compute, embedding storage, and analytics within a single environment. It integrates with internal systems such as SAP, document repositories, and RPA tools to cover the entire cycle, from bid receipt to final reporting.

The design is responsive and accessible from multiple devices, with natural-language queries that simplify access to complex information. It includes real-time monitoring to identify incidents, bottlenecks, and up-to-date metrics, strengthening transparency and decision-making. Security is a core pillar: advanced encryption, access controls, and policies aligned with standards such as OWASP are employed to protect the integrity and confidentiality of supplier and financial data.