Hi everyone,
I was advised by the team at UiPath Austria to post here regarding any additional information I might need.
I’m currently preparing to write my Bachelor’s thesis, tentatively titled:
How Practical is Agentic AI in Real-World Applications?
I attended a few demos last Thursday, which I found quite fascinating. However, I’m still trying to assess the real-world applicability of Agentic AI, and I’d appreciate any insights, use cases, or resources that could help me better understand its practical value.
To give you a better idea of what I plan to cover in my thesis, here’s a rough outline of the topics I’ll be addressing:
- Introduction: Motivation, research goals, structure of the thesis
- Fundamentals: What is Agentic AI, how it differs from classical AI, key concepts like autonomy, goal pursuit, contextual understanding, and the technological basis (e.g., LLMs, planners, tools)
- Current Research: Recent developments (AutoGPT, LangChain Agents, etc.), architectures, and open challenges
- Practical Use Cases and Limitations: Automation, customer support, DevOps, research use cases; requirements for infrastructure & security; real-world limitations (hallucinations, loss of control, debugging complexity)
- Evaluation of Practicality: Criteria like reliability, cost, maintainability; analysis of examples or even a small self-built prototype if possible
- Outlook: Future developments, ethical considerations, and research gaps
- Conclusion: Summary of findings and answering the research question
I’d really appreciate any input, experiences, or resources – especially practical insights or real-world use cases – that could help me evaluate the feasibility and maturity of Agentic AI systems.
Thanks in advance! - Kevin G.