Festina Lente, latin for hurry slowly, is an age-old principle that calls for speed and thoroughness to be balanced with one another. Although the expression has its origins in the Ancient Greek σπεῦδε βραδέως ("hurry slowly"), the thought is more relevant than ever. In a world where software is developed ever faster and AI plays an increasingly large role in writing code, the challenge is not only to build quickly, but above all to develop sustainably.

Modern development teams work with short iterations, Continuous Integration and Continuous Delivery (CI/CD). New functionalities are sometimes deployed multiple times a day. AI code assistants speed up this process further by generating code, taking over repetitive programming tasks and supporting developers with complex issues. The speed is unprecedented, but precisely because of this, the importance of quality increases.

When immature or insufficiently checked code spreads quickly, errors can multiply just as fast. Production systems therefore require a verification-first approach: quality and verification must be part of the development process from the very beginning and not just take place shortly before a release. The faster development progresses, the greater the responsibility of organisations becomes to demonstrably guarantee the quality, security and reliability of their own software.

AI also helps in this: AI can generate test cases based on requirements, analyse errors by linking log files and code changes, and even improve vulnerable or unstable test automation. Humans remain responsible for the most important decisions. Developers, architects and testers ultimately determine the quality standards, the acceptable risks and whether software is ready for production. AI accelerates the process, but does not take over this responsibility.

Testing is therefore not the final piece of software development, but an integral part of it. Unit tests, integration tests, regression tests, security tests and performance tests give developers the certainty that changes do not cause unwanted side effects. In AI applications, additional validations are added to this, such as assessing accuracy, robustness, bias and explainability of models. Only when quality is continuously validated can development speed actually be increased without losing the trust of users.

For organisations in the public sector, healthcare and financial services, this development is even more important. Their systems support critical processes and are subject to increasing societal and legal supervision. European regulations, such as the AI Act and the Digital Operational Resilience Act (DORA), turn quality assurance and testing no longer into an internal quality control, but a demonstrable part of compliance. Testing is the moment at which organisations can show that they understand how their AI-driven systems function before they are put into production.

The same principle also applies within Robotic Process Automation (RPA). Bots automate business processes, but depend on external systems, user interfaces and constantly changing business rules. Exactly a careful design and a robust test strategy ensure that automation continues to add value.

Festina Lente therefore does not mean that organisations should slow down AI-driven software development, and certainly not that they should embrace new technology without a critical eye. It means that the discipline around software quality must grow along with the speed of development. As AI generates more code and releases succeed each other faster, quality controls, automated tests and governance must also develop further. The speed of development requires equally strong safeguards.

The age-old wisdom behind Festina Lente therefore remains surprisingly relevant. Those who develop carefully, test thoroughly and consciously manage risks, precisely create the conditions to innovate faster and more reliably. Or, as software engineers often put it: "Slow is smooth, and smooth is fast." By first safeguarding quality, the speed that organisations need to be able to innovate sustainably and maintain the trust of users is ultimately created.

For more questions about Testing, send a message to: info@mvrdw.nl

By our editors.