By leveraging AI-powered agents integrated into the MGPT system, Sicredi automated critical stages of its third-party hiring process, such as creating job descriptions and screening résumés. This reduced operational effort while increasing the speed and accuracy of candidate selection.
The Third-Party Workforce Management Module (MGPT) is an internal platform developed by Sicredi in partnership with SoftDesign. Since its launch in 2018, the platform has undergone several evolutions and now centralizes the entire third-party talent lifecycle, from job requisitions and candidate screening to hiring and maintaining professional histories.
Currently, the sourcing team manages over 1,300 third-party professionals and 20 vendors through the MGPT platform. Due to rapid advancements in artificial intelligence, the team identified a strategic opportunity to integrate AI agents into the MGPT platform to streamline operations and enhance candidate selection decisions.
Two key pain points were identified in the talent acquisition workflow:
These inefficiencies impacted productivity, increased time-to-fill, raised costs, and added an unnecessary operational burden.
We developed AI agents integrated with MGPT to automate time-consuming steps in the selection process, such as creating job postings and analyzing candidate résumés.
Automated job description generation
Based on parameters provided by the hiring manager, such as seniority level, required technologies, and candidate profile, the AI generates a detailed, standardized job description. This ensures that all vendors receive complete and consistent information, increasing alignment between candidates and role requirements.
AI-powered résumé matching
Once candidates are submitted, the AI assesses their résumés using a matrix of technical and behavioral requirements and produces a compatibility ranking. This empowers hiring managers to make more informed decisions while retaining human oversight.
Our team led the opportunity mapping and technical implementation of these AI solutions, leveraging Large Language Models (LLMs) that are securely hosted within Sicredi’s Azure environment. The system architecture includes prompt engineering and retrieval-augmented generation (RAG) for contextual grounding, as well as Kafka-based queuing for efficient data processing.
Combining technical expertise, business insight, and practical sourcing experience, we delivered a production-ready solution that continues to evolve based on real-world usage data.
AI doesn’t replace human decision-making in this process; it improves it by providing better information to support the hiring decision.”
— Rodrigo Cunha, Product Owner at Sicredi.
AI-powered solutions replaced manual, repetitive steps, increased standardization, and accelerated hiring timelines.
Key outcomes from integrating LLMs into the MGPT platform that have already been observed include:
The MGPT platform continues to evolve. Upcoming developments include gathering direct user feedback on AI-generated job descriptions and tracking key metrics, such as time-to-hire and replacement rates, to measure the impact of AI.
These insights will continuously refine the AI models. Through proof-of-concept initiatives, Sicredi plans to expand AI usage to other MGPT features and stages, reducing the manual workload even more and boosting the efficiency of third-party hiring.
At SoftDesign, we believe that innovation and digital transformation should be seamless. Interested in exploring how we can accelerate your digital journey? Let’s talk. We are passionate about crafting solutions that deliver results.