Praveen Akkiraju w/ Michael Krigsman @ CXOTalk - 853

#

If you are interested in learning more about agentic AI, or you have any of these questions then without answer of fuzzy answer then this interview is for you.

Interview Questions

#

1. What is agentic AI? Please give us a background of AI as it relates to this and why agentic AI has suddenly erupted in popularity recently.
2. When we talk about agentic AI, agents, intelligent tasks, and intelligent workflows, what are we discussing?
3. How is this different from horrible chatbots that enterprise software companies sell as “great customer experience”?
4. When we talk about autonomous, can you elaborate on that? To some degree, that term separates agentic AI from traditional chatbots.
5. Is the reasoning step the crucial aspect that distinguishes agents from other types of computer science problems or applications?
6. Can we call these AI agents instead of trying to make up a new word?
7. What percentage of agentic work is focused on enterprise back-office processes? Is it more fruitful than front-end human interaction tasks?
8. In the current generation, are the agentic AI layers all based on neural networks or LLM models, or are some of them more traditional symbolic AI kinds of code?
9. Won’t these agents replace white-collar jobs also? Are these agents the death of consultants?
10. What are the potential harms that may come from automating decision-making using agentic AI at scale?
11. Can you describe how we create agents and maybe some of the companies out there who are working on this?
12. Can you talk about the right type of data that needs to be collected to implement these agents?
13. Should the agents be transparent on how they came to their conclusions?
14. Can you help technology leaders see through the agentic hype? When evaluating these technologies, what should they look for?
15. How do you ensure that agentic AI doesn’t go off the rails over time?
16. Do enterprise business architects have a role in AI, and should they report directly to the CEO?
17. Are these agents solely focused on optimizing tasks, or do they incorporate guardrails in their design?
18. What is the ultimate goal of these agents? Is it to replace people? Should employees train these agents so they won’t have a job anymore?
19. What types of jobs are most likely to be replaced by agentic AI, and which jobs will be augmented instead?
20. How do models like GPT-o1 and Google’s self-correcting reinforcement learning improve reasoning and planning for agentic AI?
21. How do concepts like chain-of-thought reasoning and reflection loops contribute to improving agentic AI?
22. What specific applications or use cases highlight the strengths of agentic AI in enterprise settings?
23. What are the challenges associated with trust and safety in deploying agentic AI at scale?
24. How do AI agents differ from traditional applications or systems of record in terms of architecture and functionality?
25. What role does the data layer play in enabling the functionality and accuracy of agentic AI?
26. How does RAG (Retrieval Augmented Generation) help enhance model performance without retraining?
27. What are the benefits and trade-offs of using open-source frameworks like LangChain or BabyAGI for building agentic AI?
28. How do enterprises balance non-deterministic outputs from LLMs with the need for deterministic outcomes in applications?

Full Interview - What is Agentic AI? Autonomous Agents and Intelligent Workflows