Roundtable Agentic AI

Roundtable Agentic AI

by ITWELT

Humanities, Social Sciences & Thought AI — News & Applications UYComputer scienceUYQArtificial intelligence (AI)
🎙 ITWELT 👥 483 📅 March 10, 2026 ⏱ 90 min 👁 11K 📄 expert opinion 🧭 2026-08-16
Available in: English (current) Français

Keywords

Agentic AIAutonomyAI agentsEnterprise AIAI security

Summary

This roundtable discussion, moderated by Christof Baumgartner, brings together four experts from different IT sectors to explore the concept of Agentic AI. Alexander Penev (Bytesource) defines it as a shift from passive to active AI, emphasizing autonomy and the ability to interact with other programs. Udo Schneider (Trend Micro) highlights the importance of long-term memory and self-optimization, distinguishing between simple AI and true agents. Amel Hankic (Aptean) contrasts agentic AI with traditional automation, noting that instead of predefined rules, agents are given goals and tools to find their own solutions. Michael Kommenda (Tieto) adds that agentic AI reduces the need for detailed specifications, allowing for more flexible execution. The discussion covers practical applications, such as automated event searching and customer inquiry drafting, as well as significant challenges including security, data governance, and the need for human oversight. The experts stress the importance of integrating AI deeply into existing systems, like ERP, and maintaining guardrails to prevent unintended actions. They also address the hype and fear surrounding AI, advocating for a balanced approach that leverages its potential while ensuring robust security and compliance. The conversation concludes with insights on customer expectations and the necessity of hands-on experimentation to understand AI’s capabilities.

200 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information is high, as the experts provide concrete examples of agentic AI implementations, such as automated event searching and customer inquiry drafting, and discuss real-world challenges like security and data governance. The argumentation is solid, with each expert offering distinct perspectives that complement each other. They acknowledge both the potential and the risks, avoiding one-sided hype. The discussion is grounded in practical experience, making it credible and useful for businesses considering agentic AI adoption.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is moderate; the discussion is based on professional experience rather than formal research. No specific sources are cited, but the experts reference industry practices and regulatory frameworks like the AI Act. The title accurately reflects the content, which is a roundtable on agentic AI. The quality of sources is acceptable for an expert panel, though it lacks external references. The title-content alignment is strong, with the discussion directly addressing the topic.

167 words

Title / Content Match

The title 'Roundtable Agentic AI' accurately reflects the content, which is a panel discussion on agentic AI, its definition, applications, and challenges.

Quality & Reliability

7/10

The roundtable features four industry experts with substantial practical experience in AI, cloud, security, and ERP systems. They provide concrete examples and balanced views on opportunities and risks. However, the discussion is largely opinion-based without rigorous citations or data, and the promotional nature of the participants' companies slightly reduces objectivity.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • AI Act — Mentioned as a regulatory framework that companies must comply with when deploying AI agents.

Concurring Sources

  • Gartner Predicts 40% of Enterprise Applications Will Feature Agentic AI by 2027 — Aligns with the experts' view that agentic AI is a growing trend in enterprise applications.

Dissenting Sources

  • AI agents are a security nightmare — This article highlights significant security risks associated with AI agents, contrasting with the experts' more optimistic view that many security issues are already solved.

Contribution & Novelties

The roundtable provides a multi-perspective view on agentic AI, combining insights from cloud consulting, cybersecurity, ERP, and AI consulting. It offers practical advice on implementation, such as the importance of human-in-the-loop and deep integration with existing systems. The discussion also demystifies the hype, emphasizing that many security challenges are already solved if proper practices are followed.

Pour aller plus loin :

  • Agentic AI — Provides a general overview of the concept.
  • AI Act — The EU regulation governing AI, relevant for compliance.
  • Model Context Protocol (MCP) — A protocol for connecting AI models to external tools, mentioned indirectly in the discussion.

101 words

Radar Profile

The radar profile shows balanced scores across all dimensions, with slightly higher scores in information quantity and quality, reflecting the experts' practical insights. The technical level is moderate, suitable for a general business audience, while reliability is solid due to the experts' experience.

Reliability 7/10

💬 No comments were provided for analysis.