Post by Titir Pal

Recently a colleague (thanks Gunjan Bhatt) brought my attention to Agentic AI. This is relatively new, and also the next evolution in AI applications. For those unfamiliar, Agentic AI are AI systems that doesn't just execute specific tasks but has the ability to autonomously execute complex, multi-stage “jobs” like testing an entire video game, developing a website from scratch (vs. creating specific content), or running a complete development project (vs. debugging a piece of code). These jobs are not only multi-staged, but also involve learning, decision making and acting autonomously. Typically Agentic AI systems are able to achieve this by having multiple “task - agents” being “managed” by a “supervisor-agent(s)”  that learns and decides which task-agent to call at different steps and situations. It's like having an intelligent, proactive digital assistant that not only follows your lead but also anticipates needs, offering creative solutions that go beyond initial requests. This article by Bernard Marr is a wonderful introduction to Agentic, its capabilities and use cases. https://lnkd.in/gpMDQFPg And the one below is very good in describing many live use cases across domains.  https://lnkd.in/gBi7hG5B These developments show how AI is enhancing workflows and human capabilities across various fields. When you think of activities done in banking operations workflows, analytics projects, or technology development, this advancement has enormous applications. It's something I suspect many of us will be developing or using in the not-too-distant future. #AgenticAI #TechTrends #DigitalTransformation