PROOFPOINT PROTECT 2025, NASHVILLE – The cybersecurity industry is facing one of its most profound shifts since the rise of the cloud. At this year’s Proofpoint Protect conference, one concept dominated conversations above all others: the agentic workspace.
Just as the digital workspace has transformed business operations over the past two decades, the agentic workspace represents a fundamentally new way of working—one where people and AI agents collaborate side by side.
While this evolution holds immense promise for productivity and innovation, it also introduces a new risk landscape. So, as organizations globally rush to adopt AI-driven tools, security teams are grappling with how to safeguard a market that’s set to grow at lightning speed.
This week, Expert Insights has been chatting to experts at the conference about the impact of AI on the workplace and where exactly the market is heading.
Introducing The Agentic Workspace
The digital workspace was defined by SaaS applications, cloud services, and collaboration platforms. It gave businesses speed and scale—but it also introduced new vulnerabilities, as attackers shifted their focus from servers and infrastructure to accounts, applications, and data. So, security had to shift, placing a stronger focus on people than perimeters.
Now, AI is reshaping the workplace once again. The agentic workspace is an environment where people don’t just use applications and digital tools; they work alongside AI assistants and agents. These agents are embedded into workflows, analyzing information, managing projects, and even executing tasks.
But while many teams are embracing AI—68% of CISOs plan to deploy AI-powered security capabilities to reduce human error and defend against advanced threats—, 54% of CISOs believe generative AI poses a security risk to their organization.
This conflict between opportunity and scepticism is something that has been echoed across the industry, and it was a hot topic at this week’s conference. On the one hand, defenders can finally gain visibility into where data resides, who has access to it, and how it’s being used in real time; on the other, attackers can scale their operations like never before.
Just like people, AI systems can be tricked, misled, or compromised, and they can fall prey to social engineering, malicious prompts, or data exfiltration attempts.
On top of this, attackers can access (and abuse) the very same GenAI tools that everyone else can. We’ve seen multiple examples of attackers using GenAI to create realistic phishing emails, fake login portals, or malicious code at scale. They’re also frequently embedding hidden prompts in innocuous-looking documents or web pages in order to fool AI systems into performing harmful actions.
So, just as security previously had to shift to accommodate the digital workspace, it needs to shift again to protect the agentic workspace.
To do this, security tools must extend beyond people and “safeguard AI agents and the points where they collaborate and share data,” says Sumit Dhawan, Proofpoint CEO.
Innovating For The Agentic Era
One session this week highlighted a striking prediction: “In the next 18 months, interactions between AI agents will overtake the number of human interactions.”
To get ahead of this, thought leaders at Proofpoint are suggesting that the industry needs to embrace a platform-based approach to AI security that encompasses four key areas:
- Authenticating interactions across humans, applications, and agents. Every message, transaction, or command must be verified.
- Uncovering and protecting against multi-channel threats. Attacks no longer happen in isolation—they flow across email, chat, cloud apps, and AI platforms.
- Building human resilience. Employees still play a critical role, but they need training to safely work alongside agents.
- Leveraging agents as defenders. AI isn’t just part of the problem; it can be part of the solution.
And underscoring this is the fact that AI agents themselves will become force multipliers for security teams: they can process vast volumes of telemetry, detect subtle anomalies, and accelerate incident response in ways that people alone cannot.
A Market Set To Boom
Proofpoint sees the AI security market as fundamentally different from past cybersecurity waves: because of the level of trust involved, large enterprises are unlikely to want to rely on unproven startups when it comes to protecting their sensitive data and critical systems.
As a result, we can expect to see a high barrier to entry for innovators hoping to break into the “security for AI” or “AI Detection and Response” (AIDR) space, as enterprises and mid-market customers consolidate their vendor relationships, creating an opportunity for established providers to dominate.
However, startups may still find success amongst smaller organizations, where adoption cycles are faster and vendor trust is easier to build.
The Road Ahead
The agentic workspace is not a distant concept—it’s here today. And with it comes a booming market, a desire to innovate at pace, and a need for collaboration.
Because ultimately, securing the agentic workspace will require more than just new tools—it’ll require security vendors, enterprises, policy makers, and researchers to work together to define new standards of trust and protection.
