ANII
AI-Powered Assistant for Forms and Technical Training
We developed an AI-powered assistant for ANII to streamline the application process by automating form validation, publishable summary generation, and technical training for internal teams.
- Capabilities
- Topics
- Regions
- 🇺🇾 Uruguay
- Published
- July 2026
The Company
Established in 2006, the National Agency for Research and Innovation (ANII) is Uruguay’s public agency responsible for promoting research, technological development, innovation, and entrepreneurship. Through funding programs, scholarships, and innovation initiatives, ANII supports projects that strengthen the country’s scientific, technological, and productive ecosystem by connecting government, academia, and industry. As a key driver of Uruguay’s national innovation strategy, ANII continues to advance its digital capabilities to streamline grant management, strengthen collaboration across the innovation ecosystem, and improve access to digital services and information. The agency continuously invests in modernizing its platforms to foster innovation, transparency, and knowledge-driven development.
The Challenge
The application forms were lengthy, highly technical, and prone to frequent submission errors. Incorrect or incomplete documentation, the manual drafting of publishable summaries, and the overall complexity of form management created friction for both applicants and internal teams.
This resulted in delays, rework, and a significant operational burden, affecting the efficiency of the evaluation process and the overall experience of those participating in the calls for proposals.
The Solution
Baufest designed and implemented an AI-powered solution composed of two integrated components. The first was an intelligent application assistant capable of automatically validating form fields, verifying supporting documents in real time, generating the Publishable Summary using generative AI, and answering questions through a chatbot powered by ANII’s official documentation.
In parallel, Baufest delivered a comprehensive technical training program for internal teams, covering the fundamentals of generative AI and Large Language Model (LLM) development, as well as advanced topics such as production architectures, Retrieval-Augmented Generation (RAG), multi-agent systems, structured outputs with Pydantic, and observability with LangSmith. The initiative accelerated AI adoption while strengthening the organization’s technical capabilities.
Benefits
- Detected and corrected application errors before submission, reducing rework and review time.
- Automated the generation of the Publishable Summary, eliminating a manual and repetitive task.
- Reduced the workload of the support team through an AI-powered chatbot.
- Enabled real-time validation of supporting documents, improving application quality.
- Delivered 20 technical training sessions covering LLMs, RAG, multi-agent systems, and production AI architectures.
- Empowered internal teams to operate, maintain, and evolve AI solutions independently, reducing reliance on external vendors.