AI Agents Are Transforming IT Operations — But At What Cost?

The enterprise IT landscape is fracturing under the weight of its own complexity. As artificial intelligence agents flood corporate environments, they’re not just adding another layer of technology — they’re exposing the fundamental brokenness of how modern IT operations actually work.

DJ Sampath, Cisco’s Senior Vice President of AI Software and Platform, has a front-row seat to this transformation. In a candid conversation with VentureBeat, he laid bare why traditional IT operations models are collapsing and why a radical new approach called AgenticOps isn’t just beneficial — it’s existentially necessary.

The Breaking Point Nobody Wants to Admit

The core problem plaguing enterprise IT today isn’t just complexity — it’s fragmentation so severe that troubleshooting has become a scavenger hunt across disconnected systems.

“A lot of times inside of these enterprises, data is sitting across multiple different silos,” Sampath explains. “For an operator to come in and start troubleshooting something, they have to go through many different dashboards, many different products, and that results in an increasing amount of time spent trying to figure out what is where before they can actually get to the root cause of an issue.”

This isn’t a minor inconvenience. It’s a productivity black hole that’s about to get catastrophically worse.

Here’s the terrifying reality: every single employee in your organization will soon have at least 10 or more AI agents working on their behalf, each handling different tasks across different systems. The math is brutal — what currently takes hours will soon take days, and what takes days will become impossible.

“This problem is only going to be tenfold, if not a hundredfold worse when you start to think about what’s really happening with the inclusion of agents,” Sampath warns.

The Three Commandments of AgenticOps

To address this impending crisis, Cisco has developed AgenticOps around three non-negotiable principles that Sampath believes must be true for any organization hoping to survive the AI agent revolution.

1. Unified Data Access: The Death of Data Silos

The first principle is brutally simple: bring all your data together or die trying. Network data, security data, application data, infrastructure data — it all needs to flow into a single, coherent system where AI agents can actually make sense of it.

“Bringing all of that stuff together is going to be incredibly important so that the agents that you are deploying to do work on your behalf can seamlessly connect the dots across the board,” Sampath says.

This isn’t just about convenience. It’s about survival. When an AI agent needs to diagnose a network issue that’s affecting application performance and potentially creating security vulnerabilities, it can’t afford to bounce between three different dashboards run by three different teams who all speak different technical languages.

2. Multiplayer-First Design: Collaboration or Catastrophe

The second principle recognizes a fundamental truth: IT operations, security operations, and network operations teams don’t work in isolation — they just pretend to.

“When you bring the IT ops person, the SecOps person, the NetOps person all together, you can troubleshoot and debug issues a whole lot faster than if you’re working in silos and copy pasting things back and forth,” Sampath explains. “It’s humans and agents working together in a synchronous environment.”

This multiplayer approach isn’t just about human collaboration. It’s about creating a shared workspace where humans and AI agents can collaborate in real-time, with context that doesn’t get lost in translation or scattered across email threads, Slack messages, and ticket systems.

3. Purpose-Built AI Models: Generic AI Is Dead

The third principle is perhaps the most counterintuitive: general-purpose AI models are fundamentally inadequate for enterprise operations.

“When you start to go into specializations, it becomes really important for these models to understand very specific things like network configuration or thread models that you care about and needs to be able to reason about that,” Sampath says.

This is why Cisco built the Deep Network Model — trained on over 40 years of operational data including CCIE expertise, production telemetry, and real-world troubleshooting scenarios. This isn’t ChatGPT with an IT badge. It’s a specialized intelligence engine designed to understand the nuances of enterprise operations.

Cisco’s AgenticOps Arsenal

Cisco’s approach to AgenticOps is comprehensive, spanning the entire enterprise technology stack. At the center is Cisco AI Canvas — an operations workspace that replaces the chaos of multiple dashboards with a unified, generative UI that enables natural language interaction with AI agents.

Within AI Canvas, operators can delegate complex tasks using plain English: “Find the source of that network latency issue affecting our e-commerce platform” or “Identify any security vulnerabilities in our cloud infrastructure.” The agents then pull telemetry, correlate signals, test hypotheses, and execute changes — all while maintaining human oversight.

The reasoning capabilities come from Cisco’s Deep Network Model, which has been trained on more operational data than most enterprises will generate in a decade. This purpose-built model delivers the kind of domain-specific intelligence that general-purpose models simply cannot match.

But the real power lies in Cisco’s platform breadth. Spanning campus, branch, cloud, and edge environments, Cisco’s agents can consume telemetry across the entire ecosystem at machine speed. With Model Context Protocol (MCP) servers implemented across Cisco products, agents gain standardized access to tools and data without custom integration work — a critical feature when you’re managing dozens of AI agents simultaneously.

The Collaboration Crisis Nobody Sees Coming

Here’s a dirty secret about IT troubleshooting: it’s already broken. The traditional approach involves raising tickets and piecing together fragmented information across multiple systems.

“People take screenshots. Sometimes it’s in Post-it notes,” Sampath reveals. “All of this information stays in completely different channels so it becomes really hard for somebody to start collecting them together.”

Cisco AI Canvas addresses this by giving teams one shared, real-time workspace for the work at hand — so context doesn’t get scattered across chats, tickets, and screen shares. Teams can collaborate live, escalate instantly, and contribute context (such as screenshots and notes) alongside the agent’s generated charts and graphs.

But the real power emerges when AI agents join these collaborative sessions.

“The machines are constantly learning from these human to machine interactions,” Sampath explains. “When you see that same problem happen again, you are that much faster in responding because the machines can assist you.”

This creates a virtuous cycle of continuous improvement. The agent remembers how you solved that problem last time and asks if you’d like to use the same approach. You’re able to hand over more work to the agent, and the time spent debugging gets compressed as the system learns and accelerates future responses.

Security: From Bottleneck to Accelerator

Historically, security has been the department of “no” — the function that slows down innovation and blocks AI adoption. But Sampath sees a radical shift on the horizon.

With the right guardrails in place — proper detection of personally identifiable information, prevention of prompt injection attacks, and robust data governance — security can actually become an enabler of AI adoption rather than a barrier.

“Employees have already experienced the productivity gains of tools like ChatGPT and want similar capabilities within their enterprise environments,” Sampath notes. “When organizations can detect personally identifiable information, prevent prompt injection attacks, and maintain proper data governance, they can unlock and unleash the AI adoption inside of the enterprise in a fundamentally different fashion.”

This isn’t just about allowing AI tools. It’s about creating an environment where AI can be deployed confidently at scale, accelerating innovation rather than impeding it.

The Identity Layer: The Unsung Hero of Cross-Domain Operations

Cross-domain data access presents one of the most complex challenges in AgenticOps implementation. Cisco’s strategic acquisitions, particularly Splunk, position the company to address this by unifying data across traditionally disconnected systems. But bringing data together is only half the battle.

Who has access to what data becomes vitally important.

Cisco is evolving its Duo platform beyond multi-factor authentication to serve as a comprehensive identity provider, with robust identity and access management baked into the platform from the beginning, not bolted on as an afterthought.

“We’re investing in identity as a very core pillar of how these agents are going to be able to pull data from different data sources with the right authorization in mind,” Sampath explains. “Should this agent have access to this type of data? Should you be correlating these types of data together to be able to solve a problem?”

This identity layer isn’t just about security compliance. It’s about enabling the kind of cross-domain operations that make AgenticOps actually work. Without proper identity management, your AI agents are either flying blind or creating massive security vulnerabilities.

The Human Element: Evolution, Not Obsolescence

As AI agents become more autonomous, the role of humans will evolve rather than disappear. This isn’t the AI apocalypse — it’s a fundamental shift in how work gets done.

“We’re always going to have humans in the loop,” Sampath says. “What you’re going to see is the complexity of the tasks that are being performed are going to be a lot more involved.”

Take coding as an example. Today, coding can be entirely agentic. The human role has shifted from manual coding, or even tab completion, to asking an agent to create code wholesale, and then verifying that it meets requirements before merging it into the codebase.

This pattern will repeat across IT operations. Humans will focus on higher-level decision-making while agents handle execution. Importantly, rollback capabilities ensure that even autonomous actions can be reversed if needed — because sometimes the agent gets it wrong, and you need to be able to hit the emergency brake.

The Wake-Up Call: Waiting Is the Wrong Move

For CIOs and CTOs, the message is clear and brutal: don’t wait.

“A lot of folks are in this holding pattern of waiting and watching,” Sampath says. “They’re waiting for AI to settle down before they make some of their decisions. And I think that is the wrong way to think about this.”

The organizations that thrive in the AgenticOps era won’t be the ones that waited for perfect conditions. They’ll be the ones that partnered with the right vendors, built the right foundations, and moved while their competitors were still in analysis paralysis.

“A partnership with the right groups of people, with the right sets of vendors, is going to help you go a whole lot faster, as opposed to trying to just stay on the fence, trying to figure out what’s right and what’s wrong,” Sampath concludes.

The AI agent revolution isn’t coming. It’s already here. The only question is whether your IT operations are ready to evolve — or whether they’ll be left behind, drowning in a sea of fragmented data, disconnected systems, and AI agents that can’t communicate with each other or with the humans who need to manage them.

AgenticOps isn’t just a new operational model. It’s the difference between surviving the AI revolution and being consumed by it.

---

Tags & Viral Phrases

• AI agents are breaking traditional IT operations models
• fragmentation, data silos, and fragmented workflows
• AgenticOps is the solution
• humans and AI collaborate in real time
• create efficiency, boost security, and allow for innovative technological applications
• current enterprise IT management is fundamentally breaking
• necessary for IT operations going forward
• the core problem plaguing enterprise IT today is fragmentation
• data is sitting across multiple different silos
• increasing amount of time spent trying to figure out what is where
• root cause of an issue
• tenfold, if not a hundredfold worse
• unified data access across silos
• multiplayer-first design
• purpose-built AI models
• humans and agents working together in a synchronous environment
• specialized operations require models trained for specific domains
• replaces multiple dashboards with a generative UI
• unified collaborative experience
• natural language to delegate actions to agents
• pulling telemetry, correlating signals, testing hypotheses, and executing changes
• human-in-the-loop control
• trained on over 40 years of operational data
• domain-specific intelligence that general-purpose models cannot match
• spans campus, branch, cloud, and edge environments
• agents gain standardized access to tools and data
• People take screenshots. Sometimes it’s in Post-it notes.
• one shared, real-time workspace for the work at hand
• humans and AI agents can collaborate in real-time
• machines are constantly learning from these human to machine interactions
• security can actually become an enabler of AI adoption
• Duo platform beyond multi-factor authentication
• comprehensive identity provider
• identity and access management baked into the platform
• humans and agents working together
• evolution, not obsolescence
• humans in the loop
• complexity of the tasks that are being performed are going to be a lot more involved
• waiting for AI to settle down is the wrong move
• partnership with the right groups of people
• partnership with the right sets of vendors
• the difference between surviving the AI revolution and being consumed by it

,