AI Agents often have wonderful intentions, but poor execution.
We are seeing a pattern repeat (as often happens with technology). There is a genuine excitement about the potential of Agentic AI in how it can reduce repetitive tasks, automate processes, and support faster results. Often these objectives are offered with the best of intentions. However, we also often see these solutions being developed by inexperienced and/or untrained people. Sometimes, employees, sometimes people come to "sell" us a solution. This happened with websites. When tools for creating websites became widely available, there was a proliferation of sites created to address an immediate or perceived gap in information sharing. The result was information fragmentation, accelerated hosting costs, and inconsistent quality. Agentic AI is repeating this pattern, but with even more potential to be disruptive, harmful, and costly. We need training, governance, and a healthy dose of caution. The "citizen Developer" unicorn needs to be replaced by an "Agentic AI Developer" on an actual job description - with appropriate guardrails and skillsets. Once my current class is over, I plan to create some content/courses on building Agentic AI with intention. Not just "build because we can", but "build with purpose"