Join our daily and weekly newsletters for the latest updates and special content related to the industry’s leading AI coverage. Get more information
While more than 20 shopkeepers announced on Agent AI -based security agents, apps and platforms RSAC 2025The most visible news of the conference is a rare, encouraging trend for security leaders. For the first time in three years, the overall effectiveness of CyberScure has improved.
Scale Venture Partners (SVP) recently released 2025 CyberScurement Context Report, Which jointly improved the average effect of cybercularity protection for the first time in three years, which increased from 48 % to 61 % in 2023 this year. According to the report, “70 percent of security leaders were more secure against general phishing attacks, of which only 28 percent of firms have reported compromise.”
The SVP also found that 77 % of CISOs believe that safety of AI/ML models and data pipelines is a priority to improve their security currency by 2025, which is more than 55 % last year. In particular, in view of the arrival of the new agent AI solutions announced in the RSAC, 75 % of firms expressed their interest in taking advantage of the AI in automating the SOC investigation using AI agents to promote large quantities of security alerts to prevent security incidents.
The increase in the number of utility is not accidental. Their results were successfully strengthened by Sisos and his teams as a result of adoption of automation, and reduced the gaps that reduce the invaders in the past.
“If you do not have a full span, the invading products are about to collapse,” Eta Moore, senior director of the Cato networks security strategy, told Venture Bat during the RSAC 2025.
Agentk AI moving ahead of the platform DNA ahead of the minimalized products
The maver’s point of view explains why CyberScure is a new definition of what is needed for the least viable product for agent AI. The RSAC 2025 revealed how much adult agent AI is becoming. There is a group of shopkeepers using Agent AI as a code -based adhesive to unite code bases and apps together, and then there are people who have been on it for years, and Aging AI is the main focus of their code base and architecture.
In this latter group, the provider of cyberciction, where the agent AI is the main platform of their platform, and in many cases, the agent continues to double its R&D costs on mastering AI. This includes Cato networks’ sauce cloud platformFor, for, for,. Cisco AI DefenseFalcon Single Agent architecture of the crowded strike, Cyber AI loop of the Dark TraceFor, for, for,. Flexible Ai Assistant of FlexibleFor, for, for,. Microsoft’s Security Caplot and Defnder XDR Sweet, Paulo Allto Networks’ cortex xsiamFor, for, for,. Sentinelon’s single platform And Victra AI’s Cogneto Platform.
Organizations that are relying on AI-driving detection with automatic containers are eliminating housing hours 40 %. They are too Almost double twice according to the possibility Fishing intervention ineffective before the background movements. The shopkeepers often rely on the identity and access management scenario on the floor to show how their agents can help trim the workload for AIwork Floose Security Operations Center (SOC) analysts.
“Identification is going to be an important factor of AI throughout the period of his life. AI agents need to be identified. They need to understand zero confidence, and how will we confirm with them,” said Vasu Jail, Vasu Jail, Vasu Jail, Vasu Jail, Vasu Jail, Vasu Jackal, Vice President of Security for Microsoft Security. It is important that we develop our safety procedures as fast as we prepare the AI. “
A joint topic of every agent AI Demo across the show floor was confusing the attack data, which rapidly gained insights in the form of a trader’s use, and then explained the containment strategy in real time.
The crowded strike showed how to find out how to hire agent AI from North Korea’s threat campaign to real -time action, to finding agent AI to real -time action. Direct Demo followed the Tradcript DPRK’s famous Cholima Since he hired a remote D -Opse services, HR checks slipped in the past and took advantage of legitimate tools, including RMM software and VS code to quietly eliminate data, including RMM software and VS code. It was a sharp reminder that the powerful, agent AI still relies on a man in the loop to find the adaptive threats and the Fine Tone model before the signal was lost in the noise of the signal.
General AI Purpose: Discovering and killing the national state’s traderkraft
These are the attacks that no one, the company, or the nation sees to come, which is most destructive and difficult to challenge and overcome. Threats were thought that they could easily dominate the minds of many bright and sophisticated technologies in cybercularity of power grid, payment, banking, or supply chain system.
Cisco Chief Product Officer Jetu Patel emphasized the urgent need for strengthening cybercare with AI so that the threats that could be disastrous when the dynamic can be disastrous and now neutralizes. “AI is basically changing everything, and CyberSocracy is at its center,” Patel said during his key. We are no longer dealing with human dangers. These attacks are taking place on a machine scale. “
Patel emphasized that AI -powered models are not contradictory: “They will not give you the same answer every time, introducing extraordinary dangers.”
CISO needs to understand today’s complex dangers and dangers
“This is not another AI thing, I promise,” said George Kurtz, CEO of the crowded strike. “I was asked to give one, and I said, ‘How do we talk about something that is really important right now, such as getting a seat on the board table?” “This punch line offered two things at the same time: Comedy Relief and a clear issue of cybersecurity leadership in 2025.
In his keynote note, “Siso’s leader to make the board seat secure,” Kurtz issued a clear call To processEven: “CyberScureti is no longer proposed to comply. It is a mandate of governance. Boards are not just getting ready. They are being forced to calculate a cyber risk as a primary business threat.
Kurtz supported his argument with severe numbers: 72 % of boards say they are actively looking for cybersecurity skills, but only 29 % have it. Kurtz said, “This is not just a talent difference. He encouraged the audience,” If you are ready to take steps, this is an opportunity. “
For the boardroom to reach the board room it had its roadmap strategy and hand:
- Keep the flow of your business at the level. “Understand where the business price arises. If you can’t speak a margin, RR, or legal risk, you will not be able to run for long.”
- Speak the language of the board. “Each board room runs on three priorities: time, money and legal risk. If you can’t translate cyber into them, you will stay on this occasion.”
- Make your brand out of the security bubble. “The members of the board are on several boards. The route is through confidence and credibility, not just technical virtues.”
Kurtz discovered the route from regulatory reforms to the impact of the board room that in 2002, Serbian Oxley converted CFOs into solid board room partners. He argued that the SEC’s 2024 violation reporting mandate does the same for Sisos. “The threat of threats, and the code of conduct, operates,” he said. “This is our moment.”
His advice was not summarized. He called on the CISOS to study proxy statements, identify the committee level requirements and identify the network with board members who “always look for the role to fill the role.” It pointed to the crowded strike Ciso Adam Zolar, now as a model on the Board of Edont Health. Zolar, Kurtz says, is the person who got his seat in the room, learning how the board runs and is seen more than the security expert.
Kurtz closed with a challenge: “I hope I will come back in ten years, still red -haired, and 50 % of boards like CFOs will see Sisos on boards. The board room is not waiting for permission. The only question is will you be?”
“Ai is not magic – it’s math”
Diana Kelly, CTO Protect AIWith a two -sided message in RSAC 2025, one of the early crowds pulled out: “Ai is not magic – this is math. And as we protect the software, we should strictly preserve the AI Life Cycle.” Its key note provided a stable background that was cut through General AI hype, highlighting the real risks of AI models, which each organization needed to defend before starting any work on its models. Kelly provided deep insights on the model poison pollution, instant injection and fraud, demanding a full stack approach for AI security.
It introduced the OSP Top 10 for General AI, emphasizing the need to protect the AI from day Zero, as partners in the initial, dangerous model with the CISO, aggressively and its treatment, results and agent chains as privileged.
Paulo Alto Networks announced Safety is intended to acquire AI The same day Kelly’s presentation, another factor driving A lot of talks about its keynote note.
RSAC 2025 suggests why the agent AI’s results are time to supply
RSAC 2025 made it clear: AI agents are entering security workflows, but boards want proof of their work. For the CISO under pressure to justify costs and reduce the risk, the focus innovation is moving towards operational effects from the hype. Real wins, including 40 % less residential time and phishing flexibility reach 70 %, the platform was achieved by stability and automatic warning discomfort, which are all proven technologies and techniques. The moment of agent AI is here, especially for shopkeepers who only enter the market.
