Gmelius Report: The State of AI Agents in 2025

Since the launch of ChatGPT in 2022, it has become increasingly clear that generative AI and autonomous systems are set to transform the way we work and live. AI agents represent the next stage in this evolution: autonomous, interconnected systems capable of making decisions and taking actions independently.

“License to Automate” is an exhaustive research report that explores the use cases and advantages of agentic AI. We also explain the underlying technology, backed by our expertise in building intelligent email solutions. 

You’ll find inspiring examples of AI agents in action, from ReliaQuest, which improves threat identification accuracy by 30% and  InnovaCare, which builds AI agents for routine healthcare processes, to our own autonomous email drafting agent. 

However, it’s also important to acknowledge the unique challenges surrounding AI agents, and we deep dive into these ethical and technical roadblocks in this report. Readers will also gain a practical understanding of the roadmap for agentic AI adoption, informed by key research findings: 

• 60 to 70% of our workday could be automated using existing technologies, including generative and agentic AI. 
• 25% of enterprises using Gen AI will also deploy agentic AI in 2025, increasing to 50% by 2027.
• 82% of organizations plan on integrating agentic AI within 1-3 years.
• 99% of the enterprise applications being built today include AI agents.
• The key use cases for AI agent adoption are research and summarization (58.2%), personal assistants and productivity (53.5%), customer service (45.8%), code generation (35.5%), and data transformation and enrichment (33.8%).
• By the end of 2025, 1 in 4 companies say that their core business processes will use AI agents—but 86% report a need to upgrade their existing infrastructure, and 42% will require a major overhaul.
• AI is expected to contribute nearly $4 trillion to global enterprise value by 2025.

Get a full copy of the report for complete insights.