Why the agentic AI era is no longer optional

What a fantastic and insightful week at Cisco Live EMEA 2026 in Amsterdam, representing boxxe alongside our customers.

For me, Cisco Live has always been about more than announcements and roadmaps. It’s about real conversations - hearing directly from Cisco’s leadership, engaging with peers across the industry, and most importantly, understanding how our customers are experiencing change on the ground.

This year, one message came through loud and clear from Jeetu Patel and the wider Cisco team:

Preparing for the agentic AI era isn’t optional, it’s essential.

There’s no shortage of AI scepticism in the market. Some still believe the AI “bubble” will eventually burst. But what we saw and heard in Amsterdam told a very different story.

Every executive I spoke with shared the same realisation: their initial expectations for AI - ambitious as they once seemed - were actually significant underestimations.

AI agents acting as digital co‑workers aren’t theoretical or futuristic ideas anymore. They’re here, now. And just because your organisation or industry hasn’t fully embraced them yet doesn’t mean others aren’t already doing so - and at speed - placing them ahead in the race we will all eventually realise we are running.

The IT landscape has already changed forever. But what does that mean for security?

Surprisingly, in many ways, not as much as you might expect.

Our core security principles still hold true. The fundamentals haven’t gone away:

Prepare → Plan → Design → Implement → Operate → Optimise

What has changed is the scale, speed and complexity at which these responsibilities must be delivered.

AI is designed to make our lives easier, but securing an AI‑enabled enterprise introduces new layers of complexity around:

• Data content and sovereignty

• Intellectual property protection

• Intent, usage and misuse

• The influence of AI hallucinations on decision‑making

These aren’t abstract or future concerns. They’re immediate, unavoidable, and already shaping security strategies today.

The good news is that securing this new era doesn’t require complexity for complexity’s sake.

A strong theme throughout Cisco Live was the power of platform‑first, unified security, with centrally managed controls and continuous innovation across the portfolio.

The value of this approach compounds quickly with tightly integrated, loosely coupled products:

Use one solution, and you’re in a strong position.
Use two or more, and the contextual intelligence between them delivers exponential value and a dramatically improved return on investment.

But achieving this requires a mindset shift.

We need to move from simply deploying firewalls to firewalling - embedding security intent everywhere, consistently.

To do this effectively, security teams must be able to:

• Enforce global policy as close to the user or workload as possible

• Gain full visibility and context across every environment

• Inspect traffic wherever it flows

• Apply dynamic, intent‑based controls consistently

• Protect the world from AI agents, and protect AI agents from the world

Whether workloads sit on‑prem, in co-location, in public cloud, within Kubernetes clusters or inside AI‑driven environments, the expectation should always be the same:

Common policy. Common enforcement. Seamless everywhere.

Cisco’s Hybrid Mesh Policy Engine, supported by continuous enhancements within Security Cloud Control, is making this vision a reality.

SASE + Unified policy = The next milestone

Another major theme of the past year which was reenforced at Cisco Live 2026 was the convergence of SSE and SD‑WAN into a unified SASE platform deeply integrated into the wider Cisco Security portfolio – and why this matters.

The promise is simple, but powerful:

One policy. Any user. Any device. Any location.

For today’s hybrid and distributed workforces, this isn’t just desirable, it’s essential.

Deeper integration between SASE and firewalling capabilities brings us closer than ever to defining a policy intent once and enforcing it across the entire enterprise, through a single dashboard informed by both dynamic constructs and placement decisions which are powered by machine learning.

This is where simplicity translates directly into operational advantage.

During the week, I had the privilege of taking part in multiple executive‑level discussions. While there’s still a journey ahead before this entire vision is fully realised, one thing is undeniable:

The pace of evolution across Cisco’s platform - particularly Cisco Security – is both undeniable and rapid.

The roadmap reinforces a message we’ve championed at boxxe for years:

A unified, platform‑driven, single‑vendor strategy is essential for long‑term organisational security and success.

Based on the conversations I had, what we saw, and what we were given a sneak peek of, I’m confident that when we return for Cisco Live EMEA in London next year, we’ll look back and say:

“The hopes of 2026 weren’t ambitions. They were predictions - and they’ve become reality.”

And, by then, we’ll already be talking about the next dream.

So, what’s next?