Leading at AI Speed: Adaptive Leadership in the Agentic Era

Why It Matters

Without upgraded infrastructure, zero‑trust security and early governance, firms risk falling behind or exposing themselves to autonomous AI failures, making the transition to agentic speed a critical strategic imperative.

Key Takeaways

• Infrastructure upgrades essential for true agentic AI speed
• Zero‑trust models must evolve to protect autonomous agents
• Governance frameworks need early integration, not post‑pilot deployment
• Data telemetry and ecosystem openness drive scalable agentic ops
• Leadership must manage hybrid human‑agent teams with new metrics

Summary

The panel discussion, titled “Leading at AI Speed: Adaptive Leadership in the Agentic Era,” examined how enterprises must redesign infrastructure, security and governance to operate at the pace of autonomous AI agents. Tim Botker of Deloitte, Jeff Schultz of Cisco and Hillary from Deloitte’s TMT practice argued that the traditional data‑center‑centric model is no longer sufficient; power, compute and especially networking must be modernized to support both cloud and edge inferencing, while “agentic ops” are required to run infrastructure at machine speed.

Key insights highlighted three major constraints: insufficient power and compute, outdated networking across data‑center, campus and branch locations, and the need to protect non‑deterministic AI agents from attacks while also preventing them from compromising corporate systems. The speakers called for a reimagined zero‑trust approach—“protecting agents from the world and protecting the world from agents”—and emphasized that safety, security and AI trust must be baked in from day one. They noted that only 39% of firms have an AI investment framework and that 97% of AI spend goes to technology, yet 70‑77% of that budget supports people, underscoring the human‑agent partnership.

Examples underscored the urgency: Cisco’s first product fully coded by AI, and Deloitte’s study showing the rapid shift of inference workloads to campus environments. The conversation also stressed that data visibility is critical—“you can’t secure what you can’t see”—requiring massive telemetry collection and cross‑ecosystem correlation. Governance must be embedded early, with lightweight risk frameworks for low‑impact pilots and stricter controls for high‑stakes use cases.

The implications are clear: leaders must adopt open, platform‑based solutions, invest in edge compute, and redesign permission models to manage agents that operate 24/7 without human consequences. Investment models are shifting toward wave‑based, outcome‑driven funding rather than large upfront caps. Ultimately, the ability to balance speed with robust controls will determine which organizations capture the competitive advantage of the agentic AI era.

Original Description

As Agentic AI accelerates faster than traditional planning and governance cycles, leaders are being forced to rethink how they decide, invest, and adapt in real time. This panel brings together senior executives to share how they are navigating this velocity gap, making strategic bets amid uncertainty, and building organizations that can learn and evolve continuously. Grounded in real-world experience, the conversation explores what it truly means to lead when adaptability is no longer a capability, but the differentiator.

Discover more about our presence at #MWC26: http://www.deloitte.com/mwc

Day 1: https://www.youtube.com/watch?v=BVCdiyIzcXw

Day 2: https://www.youtube.com/watch?v=eAJB4MBEHco

Day 3: https://www.youtube.com/watch?v=4UY0b0IO5F0