The rapid evolution of artificial intelligence (AI) has captivated industries, researchers, and enthusiasts alike. According to the 2025 Gartner Hype Cycle for Artificial Intelligence, two standout technologies are making significant strides: AI agents and AI-ready data. These technologies are hitting the Peak of Inflated Expectations this year, amid a wave of excitement, ambitious projections, and speculative promises.
Gartner’s Hype Cycles serve as visual guides, mapping the maturity and adoption of various technologies. They illustrate how these innovations can solve real business challenges and seize emerging opportunities. The Hype Cycle methodology provides insights into the technological trajectory, helping organizations align their deployments with specific business objectives.
“With AI investment remaining strong this year, a sharper emphasis is being placed on using AI for operational scalability and real-time intelligence,” shares Haritha Khandabattu, a senior director analyst at Gartner. This shift signifies a move away from merely focusing on generative AI towards foundational aspects enabling sustainable AI delivery, notably AI-ready data and AI agents.
Looking ahead, Gartner identifies multimodal AI and AI trust, risk, and security management (TRiSM) as crucial technologies poised for widespread adoption in the next five years. These developments promise to foster robust, innovative, and responsible AI applications that can significantly transform operational landscapes across businesses.
Hype Cycle for Artificial Intelligence 2025
Source: Gartner (August 2025)
“Despite the enormous potential business value of AI, it isn’t going to materialize spontaneously,” warns Khandabattu. The path to success in AI adoption hinges on closely aligning pilot projects with business needs, performing proactive infrastructure assessments, and fostering collaboration between AI builders and business teams to extract tangible value.
AI Agents
AI agents represent a new frontier in computing, characterized as autonomous or semi-autonomous software entities. They harness artificial intelligence techniques to perceive their environment, make decisions, take actions, and pursue goals—whether in digital or physical realms. Organizations are increasingly employing AI agents, utilizing advanced practices like large language models (LLMs) to tackle complex tasks.
“To fully leverage the potential of AI agents, organizations must identify the most pertinent business contexts and use cases,” emphasizes Khandabattu. However, this can be challenging since no two AI agents are alike, and their applicability often varies based on specific situational demands. As AI agents continue to evolve, understanding their limitations is essential for effective deployment.
AI-Ready Data
The backbone of successful AI applications lies in data that is optimized for AI usage, often referred to as AI-ready data. This concept focuses on ensuring that datasets meet quality and relevance criteria for specified AI applications. Evaluating data readiness is a contextual exercise; it depends on the AI use case and technique employed, forcing organizations to rethink their data management strategies.
According to Gartner, organizations that aim to scale their AI initiatives must adapt their data management practices to support current and future business needs. This includes establishing trust, mitigating risks and compliance issues, safeguarding intellectual property, and reducing biases in AI outputs.
Multimodal AI
Multimodal AI models are setting the stage for a new wave of innovation, as they are capable of learning from multiple data types—such as images, text, audio, and video—simultaneously. This comprehensive approach to data allows these models to develop a deeper understanding of complex scenarios, enhancing their ability to provide valuable insights. Over the next five years, the integration of multimodal AI will become increasingly central to the development of software applications across various sectors.
AI TRiSM
As organizations embrace AI, the importance of ethical considerations and secure deployment cannot be overlooked. This is where AI TRiSM (Trust, Risk, and Security Management) comes into play. Comprising technical capabilities that align with enterprise policies, AI TRiSM addresses governance, trustworthiness, fairness, safety, reliability, security, and data protection.
“AI introduces unique trust, risk, and security challenges that traditional controls fail to address,” comments Khandabattu. For organizations to navigate this evolving landscape, implementing layered AI TRiSM technologies is crucial for ensuring robust governance and support across all deployed AI systems.