Governance is the Top Priority for Companies Using Agentic AI Systems

Nearly 80% of IT professionals surveyed by API management firm Gravitee ranked governance as “extremely important,” underscoring the fact that while organizations are eager to innovate, they’re even more focused on doing so responsibly

/EIN News/ -- DENVER, May 01, 2025 (GLOBE NEWSWIRE) -- Gravitee, the leading Agentic API and Event Management company, announces the results of a survey on the use of Agentic AI systems and Large Language Models (LLMs) by large and midsize companies. A key takeaway from the survey is that IT professionals are aggressively adopting Agentic AI and LLMs to increase operational efficiency and improve customer experience. But many who responded to the survey expressed concern about how the use of these tools might expose their company to data privacy and security risks.

Organizations across industries are racing to harness the power of generative AI and autonomous agents. From customer service to internal automation, the promise of AI agents lies in their ability to take action, not just generate text. But as companies experiment with LLMs and agentic AI, they face challenges around governance, cost, explainability and security.

Gravitee’s study, based on responses from 300 technology leaders, developers and AI practitioners, explores the state of agentic AI and LLM adoption. It dives into how organizations are using agents and LLMs today, what they’re spending, where they’re seeing the most value, and which roadblocks are slowing broader adoption. Following are highlights from the study:

Agentic Everywhere

Survey data shows that the adoption of agentic AI is no longer experimental—it’s already mainstream. A striking 72% of respondents reported that their organizations are actively using agentic AI systems today. The data also reveals a strong pipeline of future adoption: another 21% of respondents plan to implement agentic AI systems within the next 24 months.

Why Now?

Nearly 74% of participants selected increasing operational efficiency as a leading driver for implementation, highlighting how agentic systems are being viewed as critical tools for automating repetitive tasks, reducing manual overhead and streamlining internal processes. Customer experience (46.23%) and reducing costs (37.74%) also ranked highly, signaling a shift in how organizations are applying AI, not just to innovate but to deliver bottom-line results.

It’s Not Necessarily Smooth Sailing

The survey data reveals that integration with existing systems and data privacy and security concerns top the list of challenges for companies deploying agentic systems. Meanwhile, controlling the cost of LLM interactions stands out as the single biggest concern among survey respondents, with more than twice as many respondents selecting it as the next highest issue.

If there’s one theme that cuts across all aspects of agentic AI adoption, it’s control. Nearly 76% of respondents ranked governance as “extremely important,” underscoring the fact that while organizations are eager to innovate, they’re even more focused on doing so responsibly.

How It’s Done

According to the data, the most common approach is to establish a dedicated agentic AI team, with 37.74% of respondents citing this as their primary implementation group. This trend points to the emergence of a new functional specialty within the enterprise, one that blends orchestration, prompt engineering, integration strategy and governance into a cross-disciplinary capability. Data science and engineering teams still played major roles, at 29.87% and 16.98% respectively.

How It’s Being Built

OpenAI is the dominant choice for developers and organizations looking to operationalize agents, likely due to its robust APIs, ecosystem maturity and widespread familiarity, with 48.74% of respondents selecting the platform. Google Vertex AI (10.06%), Microsoft Azure (8.81%), and IBM (8.81%) also showed strong traction. Meanwhile, OpenAI’s ChatGPT has become the de facto entry point for organizations adopting LLMs, with 86.79% of respondents indicating past usage.

How It’s Being Funded

Nearly half of respondents (49.06%) reported that their initiatives are backed by a net new budget specifically allocated for agentic AI, signaling strong executive buy-in and long-term commitment. A substantial number (35.53%) are reallocating from existing budgets, without cutting into other IT initiatives, suggesting a pragmatic “start small, prove value” approach. The largest group of respondents—49.37%—reported annual spend between $50,000 and $249,999 on LLMs, suggesting that many teams are well past the proof-of-concept stage and scaling their usage.

“The survey data mirrors what we’re hearing from customers,” said Rory Blundell, CEO of Gravitee. “Companies are anxious to implement Agentic AI systems and LLMs for the purposes of improving productivity and customer experience. But they’re being cautious too, particularly around governance and control. As companies get a better handle on how to manage these challenges, we expect Agentic AI and LLM adoption to grow even faster.”

A copy of the “2025 State of Agentic AI and LLMs” report from Gravitee can be found here.

About Gravitee

Gravitee is how you can finally manage event streams and agentic AI as securely and easily as traditional APIs – and bring together all your APIs in one place, no matter the gateway, broker or agent. Companies want to share more event- driven data than ever, but constant constraints – from security concerns, to inflexible tech, to a lack of discoverability and more – keep getting in the way. But when you can bring all your APIs together, build your API program without limits, and be ahead of where APIs are going, you can hold nothing back. That’s why 100+ global organizations rely on Gravitee, and it’s why we’re a leader in the 2024 Gartner Magic Quadrant™ for API management.

Contact:
Kevin Wolf
kevin@tgprllc.com