The Realities of Agentic AI: Bridging the Gap Between Promise and Performance

Despite the growing interest in autonomous AI, almost half of current systems remain limited to predefined actions. Organizations face significant challenges in data readiness and organizational change, hindering the full potential of agentic AI.

As the world grows increasingly enamored with artificial intelligence, the promise of autonomous AI agents remains tantalizingly out of reach for many organizations.

A recent report from the Information Services Group (ISG) reveals that, despite the hype, nearly half of the agentic AI systems currently in production are limited to executing predefined actions. This insight was unveiled at the 2025 ISG AI Impact Summit held in Boston, a gathering that underscored the gap between expectations and reality in the realm of AI capabilities.

The findings suggest that while there is a burgeoning interest in AI agents capable of autonomous operation, the majority of these systems still rely heavily on human oversight.

According to the ISG report, only 25 percent of existing solutions possess the ability to function independently. In stark contrast, a significant 45 percent of current implementations position AI as supportive advisors, necessitating human review and approval for critical workflows. This paints a picture of a technology still in its infancy, wrestling with the complexities of real-world application.

Agentic AI is touted as the next evolutionary step in artificial intelligence, aiming to enable systems that not only make complex decisions but also pursue goals with a degree of autonomy. Yet, the reality is far more nuanced.

The report highlights that many organizations struggle with legacy data issues, with over half reporting challenges in data readiness. This is a significant hurdle, especially when considering that AI does not process information in the same manner as humans or other data-driven technologies.

Olga Kupriyanova, a principal consultant at ISG and co-author of the report, emphasizes this point, stating, “While agentic AI technology is already capable of delivering enterprise value — and continues to mature, unlocking even greater potential — most enterprises are unable to realize that value due to challenges with data readiness and organizational requirements.”

This sentiment resonates with many enterprises as they grapple with the foundational changes needed to leverage AI effectively.

To bridge this gap, many organizations are turning to service providers for assistance in improving data management. These providers are increasingly employing agentic AI to address clients’ data challenges by unifying disparate sources.

However, the need for a shift in data architecture, which has long been considered the standard, remains a significant barrier.

Moreover, the report underscores the importance of organizational readiness. Companies must be willing to adapt and reimagine their processes, much like they did during the early stages of offshoring.

This readiness for change is crucial, especially as service providers are focusing on industries with the highest potential for transformation through agentic AI.

Currently, banking, financial services, and insurance dominate the use cases, accounting for nearly 30 percent. Retail and manufacturing follow closely behind, with 21 percent and 18 percent, respectively.

Interestingly, while many enterprises are still navigating the complexities of implementing agentic AI, there is already a palpable sense of excitement about the future.

Loren Absher, director of AI Advisory at ISG, notes that leading enterprises are looking beyond mere cost efficiency. They anticipate that agentic AI will not only enhance profitability but also foster the reinvention of products and business models.

“We expect companies to start by adopting agentic AI in IT infrastructure and the back office, then shift to automating more complex tasks in the front office and for service delivery,” Absher predicts.

The ISG State of Agentic AI Market report draws from extensive surveys and interviews, highlighting a comprehensive view of the landscape.

It serves as a reminder that while the technology is on the brink of remarkable advancements, substantial groundwork must be laid for organizations to reap the benefits of agentic AI fully.

As we look ahead, the evolution of agentic AI will undoubtedly continue to shape the business landscape. While the journey toward fully autonomous AI agents may be fraught with challenges, the potential rewards for those who navigate this transition successfully are immense.

With the right combination of data governance, organizational change, and a willingness to embrace innovation, enterprises can position themselves to harness the transformative power of agentic AI, paving the way for a future where AI truly augments human capabilities in unprecedented ways.