Governance and Security of Enterprise AI Take Center Stage at ”Multi-Agent Systems” Session in ”Cairo ICT 2026”
The "Multi-Agent AI Systems" session, held as part of the "Cairo ICT 2026" exhibition and conference, explored the promising future of enterprise AI with the participation of leading technology experts and decision-makers from across the region.
The session was moderated by Dr. Ahmed Abdel Latif, an international telecommunications consultant, who kicked off the discussion by highlighting a key legal development: Egypt’s recently enacted Personal Data Protection Law will apply strictly to both individuals and AI models alike.
Collaboration and Governance: Pillars of Digital Transformation
Enaith Govender, Head of Enterprise AI at Liquid C2, emphasized that building multi-agent systems is a collaborative end-to-end process that extends beyond technology alone, requiring continuous workforce upskilling alongside robust security and governance frameworks.
Meanwhile, Ahmed Rashad, AI Business Development Manager for MEA at NUTANIX, noted that security requirements become far more complex when deploying multi-agent models. Because these systems can make sophisticated decisions and reach advanced outcomes autonomously, organizations must carefully evaluate all variables before adopting them.
Risks and Oversight: Protecting Data from Unintended Execution
Mohamed Afifi, General Manager of FEDIS, warned of the risks of inadequate oversight, stating that the greatest danger lies in an AI model executing tasks it was never assigned to perform. He added that regulatory frameworks are currently being established to ensure data safety within these ecosystems.
Along the same lines, Moataz Shawky, Regional Pre-Sales Director (North Africa & Levant) at VMware - MBcom Technologies, pointed out that success is measured not merely by adopting AI, but by the operational capability to run and protect these models effectively.
A 3-Step Roadmap and an Integrated Business Model
Karim Abdallah, Head of AI Products and Cloud Software at Ericsson, outlined a three-step roadmap for implementing multi-agent systems:
1. Establish comprehensive alignment across responsible departments.
2. Clearly divide and assign tasks among AI models.
3. Independently evaluate each stage to prevent errors from cascading across processes.
Concluding the session, Hazem El-Gbeily, Chief Data and Digital Transformation Officer at Orange, emphasized that AI is not just a technology, but an integrated enterprise business model shaped by risk management, capital allocation, security, and strategic decision-making tools.
