Agentic AI is expected to play a central role in the digital transformation of enterprise systems to AI-native stacks, says data and analytics group GlobalData.
Agentic AI holds much promise for improving the efficiency of enterprise workflows, reducing costs, and improving customer experience. Able to communicate and collaborate, AI agents are being developed for a wide range of consumer, enterprise, scientific, and industrial purposes. Highly complex environments such as industrial plants or hospitals might be served by an orchestration of multiple AI agents, each tasked with a different purpose.
GlobalData defines agentic AI as advanced AI systems that act autonomously, making decisions and taking actions with limited or no human supervision. An AI agent is a software program that interacts with its environment and collects data which it uses to perform specific tasks, answer questions, and automate processes for users.
GlobalData’s latest Strategic Intelligence report – Agentic AI – notes that the agentic AI ecosystem includes software companies that offer agentic frameworks, software platforms, and agentic tools. The list of companies is growing rapidly and includes system integrators, start-ups, and Big Tech.
For enterprises that do not have the resources to create their own AI agents, the industry has already developed many pre-packaged AI agents for a wide range of applications such as earnings analysers, video script generators, and customer profile builders. Dozens of companies are already providing industry-specific agentic AI solutions – and the number is growing.
“The agentic AI ecosystem is growing rapidly,” says Isabel Al-Dhahir, principal analyst, Strategic Intelligence at GlobalData. “However, enterprise adoption will require confidence that these tools can add demonstrable business value – a detail that remains subject to ongoing skepticism. The greater autonomy and methodical approach to reasoning, problem-solving, and decision-making should see agentic AI capable of far more than previous iterations of generative AI tools. The next step is crafting these agents for practical high-value use cases.”
The industry is also pinning high hopes that agentic DevOps can improve on the past successes of robotic process automation (RPA) to enhance continuous integration/continuous delivery (CI/CD) processes and infrastructure as code (IaC) pipelines. Agentic DevOps refers to the integration of autonomous AI agents into DevOps practices, such as code assistance, to realise systems that enhance automation, decision-making, and operational resilience.
For both developers and enterprises, agentic AI is a very different proposition.
GlobalData recommends that organisations proceed cautiously regarding how much initial autonomy is given to AI agents in DevOps. Impediments to agentic DevOps progress include hallucinations in the foundational models and how readily the existing software stack can be migrated or transformed into an AI-native design.
William Rojas, research director, Strategic Intelligence at GlobalData, adds: “Not all agentic AI projects will succeed. Many will fail as developers cultivate best practices for designing, building, testing, and validating agentic AI systems. Over time, enterprises will seek to transform their software stack into an AI-native architecture. AI agents will play a critical role in facilitating this journey.”
Rojas concludes: “Integrating agentic AI into existing processes is going to be the critical challenge – clearly, it will take time for organisations to fully embrace agentic AI. Nevertheless, agentic AI will play a front-and-centre role in transforming AI-native architecture.”