AI agents and agentic security/governance

“So, the way software is actually going to be deployed out there will change drastically in the next few years. But right now, we are still in very early innings of agents.”

“I would say that what this actually new development and new areas and trending, it’s very exciting for us as overall, a B2B SaaS provider. In that I really think in the next 3 years to 5 years, fundamentally, the way software are being deployed, not only just developed now, everyone is using Copilot, those type of AI accelerated dev tools. It’s going to change right with agents introduction.”

“Agents will continue to evolve. There is – actually, it’s still very, very early stages. If you can’t get a good quality output from your current AI deployments, given poor data state, and make – deploying agents would just make it even worse, because there is so much more automation introduced.”

“So, this is actually a fantastic growth area. We're very excited about the results we're seeing in the field. So, we think that this is something that's going to be a theme carrying forward in the next few quarters.”

“Yes. You see -- we have already talked a lot about Agentic AI governance. That's the very hot topic. We have already done this in prior with the Power Platform governance, low-code and no-code application governance. And now we're working very closely with some of our largest global customers around Agentic governance.”

“Our objective is to bring the governance of a agentic AI into the same control plane as more traditional data protection and cloud governance tools so that governance is holistic and not siloed.”

“We have enhanced the AvePoint Confidence Platform to provide deeper visibility into agentic AI and specifically the life cycle of agents created in environments like Copilot Studio.”

“At AvePoint, we believe that organizations must treat agentic AI governance as a first-class discipline, just like data protection, identity and access management and cloud governance.”

“And lastly, from compliance to strategic governance. Governance cannot be just about meeting controls. It must be a strategic enabler.”

“The addition of agentic AI systems that can automatically execute workflows, interface across apps and architectures and make decisions within defined boundaries only exacerbates these risks.”

“Second, from enterprise governance to ecosystem governance, governance will expand beyond internal agents to partner develop build agents, federated agents across organizations and cross-cloud agent networks. We are preparing our platform for that complexity, including the recovery capabilities from potential agentic AI damage that I referenced earlier and evaluating a variety of pricing structures for customers.”

“First, from visibility to autonomous governance. Today, we provide dashboards, metrics and control planes. Next, we'll layer in more automated governance workflows such as agent risk scoring, automated remediation triggers, policy-driven agent retirement.”

“Companies that properly govern their agentic AI can confidently move faster and scale smarter. At the same time, organizations that neglect proper governance face delays, rework or regulatory issues, as I noted earlier.”

“We believe governance must be built in, not bolted on. That means embedding controls such as role-based access for agents, segregation of duties, process triggers for agent deployment or retirement and consistent life cycle management.”

“Today, as agents proliferate, the missing layer is not more AI. It is governance and operational oversight for AI. That is exactly why we built our sixth command center, AgentPulse, which provides unified visibility, governance, and operational oversight for agentic AI.”

“It also functions as a governance and control layer for agentic AI, providing the trusted data foundation that agents need to act safely and effectively in the AI era.”

“This was also validated by Gartner, which referenced AvePoint, Inc. in their latest research on how to build a strategy for M365 Copilot and agentic AI in 2026.”

“And lastly, building on AgentPulse, our new agentic AI governance and data protection features that we announced earlier this month provide customers with better insights about agent security posture and the ability to correct security problems directly in the Confidence Platform, helping them use agentic AI tools safely and efficiently.”

“AvePoint, Inc. customers can now inventory agents across their digital estate, surface usage, risk, and cost signals, monitor performance drift, and ultimately take action when needed.”

“We also see in the age of agentic AI, where there are more sophisticated agents being deployed, these agents are actually fully licensed. From a software perspective, from a licensing perspective, it looks like a virtual employee.”

“So you have seen—Microsoft released their agent governance capabilities. We actually announced our capabilities at the same time, AgentPulse, which covers multi-cloud, and particularly we are looking at agentic not only governance from a risk exposure perspective, access control, but also the cost.”

“As AI is becoming deployed more widely and evolves from assistance to autonomous agents, data access increases exponentially and data governance becomes top of mind.”

“Today, when I meet with customers and partners globally, the question is no longer what can AI do for my organization, but rather, can I trust, govern and operate AI safely and at scale.”

“the conversation has pivoted away from productivity and towards something far more important, enterprise trust in this new enormously powerful technology.”

“The real center of gravity, not surprisingly, has shifted to data, the knowledge that powers AI and fuels the next 2 layers, AI models and agentic AI.”

“weak data governance and poor data controls lead to bad decisions and security risks, in turn destroying trust. Ultimately, once trust is lost, AI doesn't scale.”

“This is critical because as AI agents operate more autonomously across enterprise productivity apps, companies truly need a trust layer so that they can scale AI adoption without losing control of data security, privacy and compliance.”

“the most commonly leveraged productivity tools today like Microsoft 365, Google Workspace, Salesforce and others were originally designed for human productivity and not autonomous AI execution.”

“It's why we're building the trust layer for AI, spanning data, governance, risk and operations so that organizations can deploy AI securely responsibly and with confidence.”

“We believe that organizations can only trust AI when they prioritize 3 things: first, precisely control what AI can access; second, govern and audit every action AI takes and finally, recover instantly when something goes wrong.”

“The resulting contextual data is an enormous competitive advantage for AvePoint and truly distinguishes us from legacy point solutions and backup first vendors.”

“This differentiation was also recently validated by Gartner, who specifically cited AvePoint's comprehensive set of capabilities and platform strategy as superior to native offerings like Microsoft's Agent 365.”

“New this quarter, organizations can now see across their entire agent stack, including Copilot Studio, Microsoft Foundry, SharePoint Agents and Gemini Enterprise, all within one screen in Agent Plus.”

“This quarter, we launched a new risk definition for AI agents, so organizations can better access more information about agent security and correct problems automatically.”

“This is especially critical because unmanaged agents can lead to runaway costs and expose sensitive data without proper oversight.”

“This integrated approach, see, govern and recover is powerful because it transforms AI risk into a manageable variable and ensures that the trust layer is a foundation for AI-driven growth, and it is resonating across the market, firmly cementing AvePoint as a foundational infrastructure that enables safe AI deployment at scale.”

“One of our largest customer recently shared that bundling Agent POS within the broader governance capabilities of our control suite has provided them visibility into thousands of agents without having to make a separate business case related to their MC65 deployment.”

“AI is a watershed moment for the security industry. When you look at the emergence of these frontier AI models, including models in the GPT class, they're driving two structural shifts in cybersecurity.”

“Memory pricing continues to increase; we see the trend continuing.”

“As a stand-alone business, to be a substantial part of Check Point's revenue it will likely take until 2027.”

“AI security will be both a stand-alone business and something that is embedded across our pillars.”

“At the same time, when I look at the demand side, we're seeing areas of growth. For example, we're very bullish on new AI data centers where I think we have a unique capability for the longer term.”

“It is important to note that although these revenues are still not a large portion of the total business, we see significant growing momentum for our AI security offering and that, together with CTEM, we expect to drive subscription revenue growth in the next few quarters.”

“The emerging technologies continue to perform strongly and position Check Point well to secure the rapidly growing enterprise attack surface driven by AI adoption.”

“Finally, we launched a Secure AI Advisory Service to help enterprises and government deploy and ultimately scale AI with security and responsibility.”

“Most recently, we also announced our partnership with Google Cloud. We're integrating the AI Defense Plan with Google Cloud's enterprise agent platform and this can deliver real-time runtime protection at scale.”

“We also introduced the AI Factory Security Blueprint. It's integrated with NVIDIA GPU service pinning on the server itself and provides end-to-end protection for AI infrastructure, and we are very bullish about this.”

“we launched the AI Defense Plan which is designed to secure enterprises across employee AI usage, applications and agents that both people and applications use.”

“Second, cyber attacks are undergoing structural industrialization. Agentic AI enables attackers to continuously scan global infrastructure, and they're generating a continuous flow of novel attack techniques. Manual operations are giving way to automated attack pipelines. This is what we call the Check Point AI attack factories.”

“We also see AI security impacting our services business—AI red teaming is one of the fastest-growing services.”

“As networks become agent-enabled, they create new pathways and complexities we haven't seen before.”

“The first is that the barrier to sophisticated cyber attacks is literally collapsing because AI is democratizing capabilities that were once exclusive to nation-states and a very small set of elite criminal organizations, and this is exposing a far broader set of enterprises to material risk.”

“we've even created a game called Gandalf with over one million worldwide users where thousands attack the environment every day; this feeds our small language model to improve it continuously.”

“Six months ago we acquired Lakera to help build a foundational model for security. We believe that if you want the latency, accuracy and cost profile required for enterprise security, we will need to train models tailored for security, which is a significant investment.”

“Network security is becoming more important and is one of the only places where you can prepare an organization for AI adoption. This is a multi-year effort, but I see it as a tailwind for Check Point.”

“If you believe that democratization and industrialization of attackers and agentification are real, then our firewalls are very well positioned for this future because of our ability to prevent known CVEs and deploy protections through our IPS in hours, not days or weeks.”

“Booking numbers for AI security are still not yet significant, but we see a very significant funnel that was created in recent weeks.”

“We do see, particularly for the second quarter and mainly for the second half of the year, very strong demand for our subscription packages and offerings—email, CTEM and SASE—as well as AI security.”

“We believe it will be a small part of 2026 but has significant potential for the future.”

“This includes machine identities, where we've seen robust demand and a strong and growing pipeline since the initial launch of SailPoint Machine Identity Security last fall.”

“Second, we're leading in the governance of emerging identity types, particularly machine identities and AI agents. Machine identities have proliferated, with some hidden deep within directories like Active Directory and Entra, making them difficult to track or secure.”

“When it comes to machines, other vendors primarily manage keys and certificates, not the actual access rights. Our approach goes far deeper, providing the full identity governance life cycle for the actual machines to which those keys and certificates belong.”

“AI agents represent the next frontier. These autonomous systems are now underwriting loans, onboarding customers, and executing critical operations across both cloud-native apps and legacy systems.”

“With our new offering, Agent Identity Security, planned for release this fall, we believe SailPoint is uniquely positioned to govern this new class of digital identities alongside all human and machine identities across the full spectrum of access from the cloud to the mainframe.”

“Third, we're embracing AI not just as a disruptor, but as a catalyst for cyber resilience. Just last quarter, we introduced Harbor Pilot, our AI-powered assistant, and it is already gaining strong traction.”

“Harbor Pilot embeds AI into the fabric of identity programs, servicing recommendations, guiding configurations, and driving faster, smarter decisions.”

“In many ways, Harbor Pilot functions like a digital employee with deep expertise in both SailPoint and identity security.”

“Fourth, threat prevention must evolve. And with Atlas, we are transforming how it's done. Atlas is our unified intelligence platform built on a consolidated data model and shared services architecture designed to deliver deep identity context at scale.”

“By continuously mapping identities, entitlements, behaviors, and risks, Atlas becomes the intelligent foundation of modern identity security, powering decisions with context, and fortifying the entire enterprise security stack.”

“For large enterprises, our expanded strategic alliance with Deloitte is a prime example. Together, we're enabling organizations to navigate the rise of AI agents, harnessing them as a force for efficiency while helping improve governance and security.”

“Our partnership with NVIDIA has been a game changer for enterprises embracing AI. We're taking that to the next level by co-creating a shared vision on Agentic AI. Together, NVIDIA and ServiceNow are developing out-of-the-box use cases for AI agents on the Now platform using NVIDIA NIM Agent Blueprints.”

“Now the C-suite is looking to ServiceNow to prevent another mess with AI. We do see the risk that every vendor's bots and agents will scatter like hornets fleeing the nest, and they trust ServiceNow as the governance control tower, which is a privileged position for our platform.”

“Xanadu showcased ServiceNow's leadership with Agentic AI. And until now, generative AI in the enterprise required human prompts to initiate action. We're now deploying autonomous AI agents that work with people, not just for them.”

“Using ServiceNow's single architecture platform and cross-enterprise data, ServiceNow AI agents can uniquely advance beyond pro-based activity to deep contextual comprehension.”

“We intend to be the control point that governs the deployment of Agentic AI across enterprise.”

“the way we sell that today is in a hybrid kind of pricing structure. You buy a subscription to our analysis use case usage, and then you burn it down based on the number of calls you make.”

“One agency is using ServiceNow AI agents to automate contract reviews, streamline approvals, and lower operational costs.”

“With regard to the public sector, let me be clear. We had a great Q1 because ServiceNow's strategy is rooted in government automation and modernization. This is being championed by the Trump administration with local leaders and national governments around the world following suit.”

“We are the only point vendor who has a full-blown orchestration engine with package AI agents as well as a tool to build new agents as well.”

“And critically, we view AI innovation as a continuous engine, not a one-time product release. We're accelerating our pace of AI development and expect this to be a sustained direction.”

“The security operations market is defined by three characteristics that make it structurally resilient and, in many ways, a direct beneficiary of the AI transformation happening across the enterprise.”

“The rapid advancement of AgenTik AI capabilities has created real uncertainty about the durability of many software business models, particularly those built around per-seat pricing.”

“Our expert-trained AgenTiC AI workflows are built on years of SOC expertise, trained online playbooks, and redefined through real-world analyst feedback. These are not generic models. They are purposeful engines that improve outcomes in real-time.”

“Second, detection and response. A faster discovery cycle on the attacker side means a faster response cycle on the defender side. Kenzo accelerates our MDR service from AI-assisted workflows to autonomous, machine-speed investigation.”

“In the quarter, we acquired Kenzo Security, an agentic platform built to run security operations autonomously and at machine speed. This is a direct accelerant to our AI SOC vision.”

“Data mesh shifts customers away from a per-alert investigation model to a system-driven one. Coverage scales with the environment, not headcount. This unlocks two things: a meaningful tailwind for MDR growth and a path to higher contribution margins through software-driven efficiency.”

“Rapid7 is evolving into a preemptive, agentic security platform that accelerates the entire SOC, delivered either as a managed service or a self-managed platform.”

“Combined with our SOAR capabilities and Kenzo’s agentic AI, we are moving from traditional patch management towards AI-native remediation—identifying flaws and deploying fixes autonomously.”

“You have these agents that accessing data, you need to be ahead of them in your remediation. You need to understand any abnormal behavior. If a weather forecasting agent is accessing HR records in 2 a.m. in the morning, you better know about it, and you will be amazed how often things like that are happening.”

“AI pushes existing access controls to their limits because many systems and agents inherit user access that is far too broad.”

“In the agentic world, an agent can access a huge amount of your data estate in seconds. Agents can move fast, behave unpredictably and maximize privileges by design. And if an agent doesn't have permissions, it will try to get them.”

“Connecting agents and models to data is what's blocking organizations from safely adopting AI faster. They need remediation at scale and to understand abnormal behavior, visibility alone is not enough.”

“AI-powered phishing doesn't just target humans. It targets agents too. Agents can read e-mail, Slack and key messages. One human clicking maliciously is one compromised identity. An army of agents can multiply the attack surface.”

“You need an inventory of every model, agent and pipeline running in your environment and you need access posture to know what data they can touch, what permissions they have and where they are vulnerable.”

“Only AI can defend AI risk.”

“With the acquisition of Atlas, we have the ultimate control plan for agents, models and pipeline. We protect every data type.”

“At our GitLab 17 event, we showcased GitLab Duo, a workflow and autonomous agent designed to take more initiative. This is where the industry is headed, and we are scaling our operations in alignment with this direction.”

“Additionally, we anticipate more autonomous AI and a shift in augmented reality from reactive to proactive.”

“We are also leveraging our reputation in open source and co-creation to extend our strategy into agentic AI.”

“Another advantage we have is our single database platform, which allows us to create a knowledge graph that integrates all relevant context for tasks, whether performed by a human or an agent.”

“Furthermore, we are committed to maintaining our standards for security and privacy, serving our clients with cloud neutrality and support for self-hosted models.”

“Additionally, we will offer comprehensive extensibility and flexibility through an open community approach for agent discovery and development.”

“Lastly, we are focusing on integrating software life cycle agents directly into GitLab, enabling every Premium and Ultimate customer to collaborate with them seamlessly right out of the box.”

“We believe this agentic vision of AI will be hugely impactful in security, and we're moving swiftly to adopt that into our products.”

“And we announced the intent to acquire Protect.ai, an early innovative leader in security for AI providing AI model scanning in the red teaming to further bolster our capabilities.”

“We've shared our initiatives in two main areas: what we're doing on Prisma AIRS to secure agentic systems and a preview of agentics, outlining our strategy for developing agentic platforms for our customers.”

“No one else possesses the same level of integration and automation that can effectively work alongside AI.”

“AI will be incessantly banging at those edges trying to find the loopholes and the misconfiguration in the way humans have not secured the perimeter.”

“Prisma AIRS extends our existing capability in posture management and runtime security and will add security for AI agents in the future.”

“With agentic AI, I see it as a transition from AI serving as a helper to gaining a level of autonomy, where it can take actions independently. Over time, as trust is established, these actions will be regulated and gradually expanded.”

“We recently announced Agent Rewind, built on our Rubrik's secure data platform underpinned by Predibase AI technology.”

“The Predibase platform allows enterprises to fine-tune GenAI models and run an optimized inference stack for faster accurate results at lower cost.”

“We are still in the early stages of optimizing product market fit for our AI solutions including Agent Rewind. We plan to add more capabilities and investments to enable confident enterprise AI transformation and agentic work adoption.”

“So we are helping do that cyber resilience transformation for our customers. Giving them AI-based ransomware detection, fast recovery, capabilities like that.”

“So we are looking at AI in a holistic way. But we are not just focused on securing the AI. What we are focused on is the cyber resilience business, as well as AI operations business, which is about agent fine-tuning, serving, agent rewind, plus plus.”

“With Agent Rewind, we can now help customers undo the mistakes of AI agents without full system rollback, which is crucial for scalable and secure AI adoption.”

“What started with Generative AI has quickly evolved to retrieval augmented generation or RAG, inference AI, and more recently agentic AI.”

“that's leading into a foundational approach to agentic AI, where we're really driving the integration not only into Dynatrace agents but also third-party agents to create a truly autonomous system. We think this will be a critical evolution in the observability industry that we believe that we're well poised to take advantage of.”

“We are driving toward an agentic AI ecosystem in which our AI agents will interoperate with third-party AI agents to take appropriate action. Importantly, deterministic answers are critical for agentic AI to work properly.”

“We recently announced significant advancements in our platform, evolving its ability to serve as the knowledge, reasoning, planning and actioning framework of agentic AI and provide trustworthy precision and adaptability.”

“We believe there are extensive opportunities to help customers responsibly utilize AI in their production environments, and Satori can help Commvault accelerate this vision.”

“At the Gartner Security and Risk Management Summit in June, analysts reported that by 2028, 25% of enterprise breaches will be traced back to AI agent abuse from both external and malicious internal actors.”

“As our customers start training models and implementing AI internally, having policy enforcement and observability for large language models and the training data becomes essential.”

“In contrast to conventional continuous threat exposure management tools that merely highlight vulnerabilities without sufficient remediation capabilities, our unique ETM solution integrates CRQ, CTEM, and native remediation operations to rapidly address the most critical risks at scale.”

“Next in line for our ETM solution, I am particularly excited about a new groundbreaking feature, TruConfirm. This feature harnesses the power of our platform to validate whether vulnerabilities are exploitable before customers face a breach.”

“While traditional security operations centers work to identify breaches post-occurrence, Qualys is leading the creation of the first Agentic AI Risk Operations Center, ROC, which aims to centralize an organization's threat response before any business impact occurs.”