Chilean scale-up wins fund that drives digital transformation with self-repairing bots
With a fund of one billion pesos, the Chilean company Rocketbot is leading the revolution in automation with self-repairing bots, so that the system identifies, diagnoses, and resolves errors autonomously and in real time, reducing human intervention and downtime to a minimum.
In a milestone that will boost its ability to take automation technology to the next level, Rocketbot, secured funding from the Innova High Technology program by CORFO, which has a funding of one billion pesos and with which the Chilean scale-up seeks to implement more efficient, self-repairing and self-configuring bot and robot systems.
One of the most innovative features of the project is the implementation of a self-repair mechanism for bots, what does this mean? The idea is for the system to identify, diagnose, and resolve errors autonomously. This works through a detailed analysis of execution logs, where each robot records the tasks performed and the commands executed.
When a failure occurs, the system automatically identifies the line and type of error, using a Python-based format that facilitates the accuracy of the diagnosis. Simply put, Herrera explains that " the goal is for bots to be able to overcome obstacles and adapt to changes in real time, minimizing human intervention and downtime ".
For Juan Jorge Herrera, CEO of Rocketbot, the self-repairing bots initiative not only improves the responsiveness and adaptability of our solutions, but also marks the beginning of key collaborations with universities and research centers that will be fundamental to meeting the high standards we have set for ourselves.
Follow us on: 
Challenge: Multi-agents
The errors that the system can solve include common problems such as changes in the appearance of a web page, the appearance of unexpected popups or the handling of unstructured data. However, limitations are also recognized, such as the need for human intervention when credentials fail or essential data sources.
To overcome these challenges, Rocketbot is exploring advanced multi-agent methodologies, where several systems work together to validate and optimize proposed solutions. This approach ensures redundancy and accuracy, mitigating risks even in complex scenarios. "Our goal is to achieve a zero error rate in critical tasks, ensuring that systems can iterate solutions until they find an optimal response or, ultimately, notify the human operator", details Herrera.
Academic Collaboration
In addition, the self-repairing bot project involves collaboration with the National Center for Artificial Intelligence (CENIA) and the Pontifical Catholic University of Chile, amongst other institutions, in developing an on-premise infrastructure that meets the strictest security requirements. This approach ensures that the solution can operate independent of large external infrastructures, preserving the privacy of sensitive data.
The methodology includes the anonymization of critical information, such as names and identification numbers, replacing them with generic data during processing. "We are designing an architecture that combines advanced encryption, anonymization, and specialized models, ensuring not only efficiency, but also the security of our clients in highly regulated sectors such as finance," adds Herrera.
The scale-up also plans to integrate this technology with the most advanced artificial intelligence models, such as OpenAI, Google Gemini, and Meta Llama, selecting the optimal model based on the needs of each process. The aim is to perfect the self-repair methodology through collaboration with these platforms, reducing the consumption of computational resources and improving real-time response capacity. Likewise, the development of a proprietary model based on Llama 3.2 is being explored, specifically designed for on-premise environments that prioritize security and data understanding.
In real environments, Rocketbot has already begun to validate its solution with customers such as Consorcio, who are testing secure and self-repairing automation in critical tasks. This not only ensures the viability of the technology, but also paves the way for its implementation in other industries. "We want our customers to trust that the technology not only optimizes processes, but also has the ability to adapt, learn and evolve with the business needs", emphasizes Herrera.
With this project, Rocketbot not only aims to lead automation in Chile, but also to set a precedent in the global industry, demonstrating how artificial intelligence can be integrated securely, efficiently, and autonomously to transform the way businesses operate.