Having just returned from the RSA Conference 2025, without a doubt the word on everyone’s lips and the dominant theme on every vendor stand was – you’ve guessed it – AI. AI is a phenomenon that just keeps evolving. Today analysts are predicting a $632B+ AI spend by 2028.
What was interesting is that the conversation has also evolved and moved from GenAI to SynthAI and agentic AI.
Not All AI is the Same
It is interesting how easily the different AI-related buzzwords get bandied around and are often used interchangeably. However, the reality is that GenAI, SynthAI and agentic AI are very different.
GenAI, or Generative AI: GenAI refers to artificial intelligence that can create original content, such as text, images, videos, audio, or code, based on patterns learned from vast amounts of data.
SynthAI: Contrary to GenAI, that primarily focuses on the divergence of information, generating new content based on specific instructions, SynthAI developments emphasize the convergence of information, presenting less but more pertinent content by synthesizing available data. SynthAI will enhance the quality and speed of decision-making, potentially making decisions autonomously. The most evident application lies in summarizing large volumes of information that humans would be unable to thoroughly examine and comprehend independently. SynthAI’s true value will be in aiding humans to make more informed decisions efficiently. A real world example is how SynthAI is helping Siemens accelerate AI adoption in industrial automation, robotics, and manufacturing by streamlining data generation and training processes.
Agentic AI: Agentic AI refers to autonomous AI agents that can make decisions, take actions, and adapt to new information with minimal human oversight. Unlike GenAI, which follows predefined rules, agentic AI operates dynamically, solving complex problems and executing tasks independently. When making decisions and taking action, these agents will rely on synthesizing and analyzing data to make said decisions (including SynthAI).
The AI Trust Gap
Trust in AI also needs to evolve. This isn’t a surprise as AI, like all technologies, is going through the hype cycle and in the same way that cloud and automation suffered with issues around trust in the early stages of maturity, so AI is following a very similar pattern. It will be some time before trust and confidence are in balance with AI.
The Rise of Agentic AI
Agentic AI was front and center of the conversation at RSA. This year we witnessed a flurry of announcements around agentic AI. Google announced AI-driven security agents for automated rule creation, malware analysis, and alert triage, integrating Mandiant services into its security platforms making it easier to build AI agents. These multi-agent AI systems are designed to revolutionize enterprise workflows and transform businesses. SentinelOne unveiled agentic AI functionality that mimics advanced SOC analysts, automating investigations and orchestrating multi-step threat responses. Likewise, ArmorCode launched Anya, an agentic AI solution for AppSec and product security teams, designed to reduce alert fatigue and accelerate security decision-making. This list goes on.
Agentic AI encompasses tools that can understand objectives, make decisions, and act. These tools streamline processes, automate tasks, and provide intelligent insights to aid in quick decision making. In a use case involving repetitive processes, take a call center as an example, agentic AI can have significant value. The integration of agentic AI in contact centers is transforming customer experience by enhancing service delivery, improving operational efficiency, and redefining the role of human agents.
In contact centers, AI can be used to design customer experience assistants that augment and enhance human experience. Such assistants can manage simple tasks, thereby freeing human agents to concentrate on more complex issues. Over time they can assume increasing responsibilities and eventually manage customer interactions.
But where threat intelligence gathering and analysis, threat hunting and forensic examination are concerned, this is much more of a grey area and still needs a human in the loop to validate the decision. When you are researching emerging cyber threats and vulnerabilities to stay ahead of attackers or undertaking forensic analysis to determine the root causes of an incident, there aren’t always binary answers. Adversaries are human and they are not always predictable or logical in their behavior.
Cybersecurity Applications of Agentic AI
Agentic AI can be a great tool for many of the ‘gray area’ tasks that SOC analysts undertake. For example:
- Autonomous AI will detect cyber threats faster, providing real-time defensive countermeasures against adversarial attacks.
- AI-powered SynthAI security bots will manage incident response, reducing workload for cybersecurity professionals.
- AI will generate security playbooks and execute autonomous threat mitigation strategies helping to automate SOC operations.
However, it is worth noting that there are other, softer, challenges to take into consideration when you are using AI – of whatever flavor – including privacy, unfair bias, ethics and fairness, hallucinations, alignment and control and the need to improve accountability.
At a pre-RSA panel that I attended, what was sparking conversation was whether AI could really replace level 1 SOC analysts? Based on the above, everyone agreed that we are not there yet. Successfully scaling projects require responsible AI practices, a clear roadmap and governance policies, however many companies lack structured AI strategies right now.
Companies who harness the potential of AI will certainly be able to foster innovation and build future-ready solutions. I’d be interested to understand how companies plan to implement agentic AI or SynthAI and to understand some of the real-world use cases for these tools.
Learn More at the AI Risk Summit
Related: How Hackers Manipulate Agentic AI With Prompt Engineering
Related: How Agentic AI will be Weaponized for Social Engineering Attacks
