The idea that “it is better to do a little well than a great deal badly,” often attributed to the philosopher Socrates, has taken on new relevance in the age of artificial intelligence. As AI tools like ChatGPT and Perplexity AI become more integrated into everyday work, experts are increasingly emphasizing the importance of using them in focused, well-defined ways rather than relying on them for overly broad or continuous tasks.
Recent discussions in AI research suggest that the most effective and safest use of AI comes from breaking work into smaller, clearly scoped tasks. Instead of engaging in long, unstructured sessions with AI systems, users are encouraged to assign specific problems with measurable outcomes. This approach not only improves accuracy but also makes it easier to verify the results and reduce the risk of errors or misinformation.
In professional environments, this principle is especially important as companies begin testing more advanced “agentic” AI systems capable of performing multi-step actions on their own. While these tools offer powerful automation capabilities, they also introduce new risks when used without clear boundaries or oversight. Keeping tasks narrow and well-defined helps ensure that AI remains a supportive tool rather than a source of confusion or unintended consequences.
Ultimately, the growing consensus is that productivity with AI is not about doing everything at once, but about doing each part well. By using AI for targeted, manageable tasks and maintaining human oversight throughout the process, users can achieve better outcomes while staying in control of the work they produce.