In this interview from Proofpoint Protect 2026, Molly McLain Sterling, senior director of cybersecurity strategy at Proofpoint, joins theCUBE's John Furrier to discuss how security strategies are evolving as AI agents become embedded in everyday enterprise workflows. Sterling, who came to Proofpoint as a customer, explains why organizations are shifting from caution to action, establishing guardrails and clarifying agent intent as AI takes on more autonomous tasks. She notes email remains the top attack vector, with phishing messages growing harder to spot as generative AI perfects grammar, branding and localization.

The conversation also explores how Proofpoint's "Threat Flip" process is breaking down long-standing silos between security operations and awareness teams, translating complex threat data into plain language that fuels real-world phishing simulations. Sterling warns that traditional patch cycles can no longer keep pace with AI-accelerated threats, citing projections that new vulnerabilities could surface every minute by 2027 and underscoring the need for layered defenses. She also discusses Proofpoint's Women Who Cyber program, which drew a nearly even split of women and men to its workshops last year, focused on cutting through AI-driven noise to find the signals that matter. From reskilling security teams with sharper soft skills to helping leaders reframe security as a growth enabler rather than an obstacle, Sterling outlines how organizations can build trust between humans and AI agents as autonomous systems take on more responsibility across the enterprise.

In this interview from Proofpoint Protect 2026, Sumit Dhawan, chief executive officer of Proofpoint, joins theCUBE's John Furrier to discuss how the rise of autonomous AI agents is forcing a fundamental rethink of enterprise security strategy. Dhawan points to real-world incidents, including the Hugging Face breach and a fully agent-driven attack campaign dubbed BlueMoon, to show how adversaries now run reconnaissance and social engineering with no humans directly in the loop. He explains that the attack surface has expanded beyond networks to include every person, mailbox and collaboration channel, and details Proofpoint's new intent-based defenses, an Agentic Collaboration Security System and Agentic Data and AI Security system, both powered by a shared, cross-customer knowledge graph.

The conversation explores how Proofpoint translates plain-English governance rules into enforceable, intent-based controls that determine what humans and AI agents can access across applications, data and MCP integrations. Dhawan breaks down why this "headless application" shift means AI runtime security and data security can no longer be managed separately, and how improving model economics are replacing manual SOC and threat-hunting work with agent-to-agent automation. He also touches on Proofpoint's disciplined growth strategy, anchored by acquisitions such as Normalyze and Acuvity, and its focus on becoming the platform customers turn to for securing the agentic workspace rather than competing product by product. From pushing back on industry "slow down" rhetoric to framing security as the foundation of competitive speed, Dhawan makes the case that focus, not caution, is what lets enterprises move fast securely.

In this interview from Proofpoint Protect 2026, Mayank ("MC") Choudhary, executive vice president and general manager of the Data Security and Governance Group at Proofpoint, joins theCUBE's John Furrier to discuss the architectural shift required to secure data as AI agents become first-class actors alongside humans. Choudhary explains why headless AI applications have erased the traditional permissions layer, forcing organizations to govern intent and access rather than static, human-written rules. He details Proofpoint's knowledge graph, a living system connecting every actor, action and piece of data, and unpacks how it decouples security intelligence from the underlying AI models powering agents.

The conversation also explores Proofpoint's new agentic data security platform announced at Protect 2026, built to detect, investigate and self-govern risk without manual controls. Choudhary notes that a company of just 3,000 employees can generate 25 million agent requests and half a million data exposures a year, a scale no human team can police manually. He explains how the platform extends the human insider threat model to AI agents, treating every autonomous action as an extension of human intent, and why customer data and models are processed on-premises to preserve privacy while giving enterprises the flexibility to bring their own AI. From reframing insider risk for the agentic era to establishing always-on runtime security, Choudhary makes the case that governance must evolve as fast as the AI it protects.

In this interview from Proofpoint Protect 2026, Tom Corn, executive vice president and general manager of the Threat Protection Group at Proofpoint, joins theCUBE's John Furrier to discuss how security strategies are evolving to govern AI agents and protect human-AI interactions across the enterprise. Corn explains the shift from machine learning and generative AI toward agentic AI and intent-based protections, powered by a knowledge graph spanning both attacker and defender data. He details a multi-tiered detection model — flash, extended and deep-thinking tiers — built to catch attacks that mimic legitimate behavior, including a business email compromise scheme that hijacked a real invoice thread with a compromised supplier.

The conversation also explores how Proofpoint is architecting for agentic ecosystems, including exposing MCP access through its Satori interface so AI agents, not just humans, can interact securely with its platform. Corn touches on the "Community Hyperloop," a mechanism that rapidly spreads threat learnings across the company's customer base at machine speed. He underscores the risks of shadow AI, where employees grant agents broad access to drives, passwords and financial accounts without oversight, and argues that consolidated platforms — rather than dozens of disconnected point products — are essential to closing the gaps attackers exploit. From intent-based detection models to agent-to-agent protocols, Corn outlines how Proofpoint is positioning itself at the center of the human-AI security shift.

In this interview from Proofpoint Protect 2026, Ryan Kalember, chief strategy officer at Proofpoint, joins theCUBE's John Furrier to discuss how enterprises are treating AI agents as a new class of insider threat that demands real-time governance and control. Kalember frames the shift around a simple but powerful lens: agents are insiders, and security teams should plan accordingly. He explains why the endpoint, not the cloud, has become the critical vantage point for monitoring agentic activity, noting that 99% of the agentic behavior Proofpoint observes today happens on regular endpoints. He details how tools like Claude Code and Claude Cowork have expanded what agents can autonomously acquire and execute, raising the stakes for enterprises still building governance around this new class of digital insider.

The conversation also explores Proofpoint's platform strategy, which extends its heritage in protecting people and data to include agents as a third pillar of enterprise risk. Kalember breaks down a layered security funnel — from deterministic access controls to intent-based detection to full agentic investigation — and explains how distilled models can now conduct round-the-clock insider risk investigations without the bias or fatigue of human analysts. He touches on how prompt injection has evolved into a form of social engineering aimed directly at agents, and why Proofpoint's market lead in collaboration and data security, powered by its Nexus knowledge graph, gives it a distinct data advantage for training specialized AI models. From nation-state actors chaining zero-day exploits to reimagining governance policies as language agents can act on in real time, Kalember outlines why the human-AI relationship remains the central battleground of enterprise security.

In this interview from Proofpoint Protect 2026, Robert Bair, head of national security partnerships at Anthropic, joins theCUBE's John Furrier to discuss how enterprises are governing agentic AI as it becomes embedded across everyday workflows. Bair draws parallels to the early cloud era, noting how skepticism around security once slowed adoption until trust was established. He explains why agents now carry their own identities and take autonomous actions on behalf of humans, requiring new governance structures built around visibility, compliance hooks and inference telemetry rather than static perimeter defenses.

The conversation also explores the widening human threat landscape, as increasingly capable models, including open-source variants, can chain together vulnerabilities into full exploits, raising the stakes for defenders. Bair details Anthropic's decision to delay general availability of its Mythos Preview model, giving security partners like Proofpoint early access to build safeguards before less careful actors could exploit the technology. He touches on data loss prevention challenges unique to agentic deployments, using the analogy of proprietary data being irretrievably absorbed into an LLM, and stresses the importance of sandboxing and gradual rollouts to control blast radius. From governance frameworks to communication skills as an enduring human differentiator, Bair outlines why trust, engineered deliberately at every layer, is the foundation for scaling agentic AI safely across the enterprise.

In this interview from Proofpoint Protect 2026, Remi Thomas, chief financial officer of Proofpoint, joins theCUBE's John Furrier to discuss how the company's growth strategy is evolving to secure the emerging agentic workspace. Thomas details Proofpoint's progress toward $2.5 billion in annual recurring revenue, fueled by organic growth and the integration of Hornetsecurity, a combined business now generating roughly $350 million in ARR and growing more than 20% organically. He outlines three areas of investment — collaboration security, data security and AI security — built to create a unified platform that protects both human and agent-driven interactions.

The conversation also explores how consolidation is reshaping customer budgets, with Thomas describing a large Canadian bank deal that replaced seven modules from four vendors with a single Proofpoint solution, freeing up funds customers can redirect toward AI adoption. He discusses Proofpoint's disciplined M&A philosophy, favoring tuck-in acquisitions such as Normalyze that deliver faster time to market over lengthy organic builds, while maintaining the cash flow discipline to self-fund each deal. Thomas also touches on industry inflection points, including the early-year "SaaS apocalypse" scare and the shift toward AI-native security partnerships that give CISOs confidence to deploy AI safely. From consolidating point solutions into a single human-centric platform to disciplined M&A that self-funds new AI security capabilities, Thomas offers a clear view of how Proofpoint is positioning itself as a strategic partner for the agentic era.

In this interview from Proofpoint Protect 2026, Selena Larson, principal threat researcher at Proofpoint Inc., joins theCUBE's John Furrier to discuss how AI is reshaping both sides of the cybersecurity battle, sharpening the tools of skilled attackers while exposing the sloppiness of low-level threat actors. Larson explains that while defenders are building more dynamic, behavioral detections, the threat landscape is bifurcating: elite actors are refining malware and phishing kits, while less sophisticated ones grow careless, leaving exposed infrastructure and generic lures. She details how attackers now localize phishing campaigns across more than a dozen languages simultaneously and deploy evolving CAPTCHA and filtering tricks to dodge sandboxes, pushing security teams to move from static indicators toward adaptive, real-time detection models.

The conversation also explores the emerging risks of agentic AI, with Larson drawing parallels to early social platforms where unchecked app permissions led to widespread data exposure. She warns that autonomous coding agents accepting unreviewed commits could open new software supply chain vulnerabilities, stressing that human oversight and governance remain essential even as AI-enabled security operations scale. Larson also highlights the BlueMoon exploit chain, a Chinese espionage campaign believed to use AI to reverse-engineer a Chrome patch gap into a working exploit, alongside real-world takedowns like the disruption of the Tycoon 2FA phishing service. From generational shifts in digital trust to the rise of AI-as-a-service phishing kits, Larson provides a clear-eyed view of how organizations must balance automation with accountability to stay ahead of an increasingly fragmented threat landscape.