Survey: 86% of Enterprises Require Tech Stack Upgrades to Properly Deploy AI Agents

SAN FRANCISCO, Dec. 17, 2024 — The “State of AI Agent Development Strategies in the Enterprise” survey of over 1,000 enterprise technology leaders and practitioners found that more than 86% of enterprises require upgrades to their existing tech stack in order to deploy AI agents.

The research was announced today and commissioned by Tray.ai. It reveals widespread integration challenges, with 42% of enterprises needing access to eight or more data sources to deploy AI agents successfully, and security concerns emerging as the top challenge across both leadership (53%) and practitioners (62%). This complex interplay between technical requirements and system readiness presents major challenges for enterprises looking to capitalize on the transformative potential of AI and agents.

“Enterprises are navigating an increasingly crowded sea of AI-enabled SaaS applications and confronting a perfect storm of integration complexity and organizational barriers,” said Rich Waldron, co-founder and CEO of Tray.ai. “The survey reveals an inflection point for enterprise AI adoption — while more than two-thirds of organizations expect AI agents to power more than a quarter of their core processes by 2025, they’re realizing that success requires rethinking how they handle data integration, security and accessibility across systems.”

Enterprise AI Agent Investment Signals Need for Scalable integration platform

The survey reveals plans to make significant investments in AI agents, with 42% of enterprises planning to build over 100 AI agent prototypes and 68% budgeting $500,000 or more annually on AI agent initiatives. However, this scale of deployment faces serious hurdles without a unified integration platform, as evidenced by nearly half of respondents (48%) reporting their existing integration platform as a service (iPaaS) products are only “somewhat ready” for AI’s data demands.

Integration Complexity Drives Platform Requirements

Organizations are grappling with foundational integration challenges in their AI agent initiatives, with 42% requiring eight or more connections to data sources. While 90% of enterprises view integration with organizational systems as “essential,” they are currently favoring hybrid approaches — a mix of build and buy (41%), single-purpose SaaS app agents (28%) or custom development (24%). Expecting these agent delivery alternatives to have the requisite connectivity to so many data sources and the underlying security and scale capabilities suggests a potential blind spot in how IT plans to address enterprise-wide integration needs.

“We’re seeing a concerning pattern in enterprise AI implementation that is reminiscent of the early days of cloud adoption — organizations clearly understand that seamless integration is essential, yet many are opting for patchwork approaches that will prove costly down the line. We’ve seen this story before: starting with custom builds and point solutions inevitably leads to a complex web of connections that becomes increasingly difficult to maintain and scale. With AI requiring unprecedented access to enterprise systems, now is not the time to create tomorrow’s technical debt,” continued Waldron.

AI Priorities Balance Operational Efficiency with Customer Impact

Enterprises are prioritizing AI agents that solve critical business problems; top use cases reported were IT service desk automation (61%), data processing/analytics (40%) and code development/testing (36%). Additionally, enterprises are focused on using AI to improve customer-facing processes, with 49% prioritizing increased customer satisfaction as a key success metric.

The survey also found that enterprises view AI as a way to drive greater efficiency, with 64% citing cost reduction as a top priority and 52% aiming to increase process automation rates. Despite the emphasis on productivity and efficiency gains, 24% of enterprises also see positive revenue impact as an important measure of AI agent success.

Leaders and Practitioners Share AI Agent Priorities, with Practitioners More Alert to Security and Data Governance

The research reveals both commonalities in the priorities of leadership (team lead, manager/senior manager, director/department head, president/vice president/senior vice president, C-level) and practitioners (practitioner, developer, solution architect, enterprise architect, software engineer) as well as a slight disparity in challenges with developing and deploying AI agents.

IT service desk automation is the top business problem both groups will prioritize for AI agents to solve, at 63% for leadership and 55% for practitioners. Other top priorities for both groups include streamlining specific operational workflows like code development and testing and data processing analytics.

Both leadership and practitioners identified security concerns as a top challenge they are currently facing in developing and deploying AI agents, at 53% for leadership and 62% for practitioners. Other top challenges for both groups were data governance, performance issues and integration complexity. However, practitioners prioritize security (62%) and data governance (49%) slightly higher than leadership — 53% and 40%, respectively.

While leadership and practitioners are largely aligned on most challenges, the slight disparity when it comes to security and data governance indicates an opportunity for better communication and a more unified outlook across these two groups.

Meeting strategic security, integration and process automation needs, as well as operational efficiency goals, will be important for successful AI agent deployments.

Organizations Remain Optimistic Despite Timeline Challenges

Enterprises remain committed to AI agent adoption, with top goals including improving efficiency, enhancing productivity and boosting customer satisfaction. However, the delta between currently slow versus the desired faster deployment speeds (64% want three-week deployments) emphasizes the need for more efficient implementation approaches.

“The next generation of iPaaS solutions must rise to meet the unique demands of AI. Organizations are realizing that the real challenge isn’t just deploying individual AI agents, but creating a truly AI-ready environment that can support their expanding needs in a rapidly changing tech environment,” concluded Waldron. “To truly unlock the potential of AI agents, enterprises must move beyond a fragmented approach and embrace unified, composable platforms that can break down silos, streamline complex workflows and provide a foundation for AI success at scale. Those who fail to address these challenges risk being left behind in an increasingly AI-driven business landscape.”

Survey Methodology

The Tray.ai “State of AI Agent Development Strategies in the Enterprise” survey findings are based on the results of an online survey that examined the opinions of 1,045 U.S.-based enterprise technology professionals at organizations with 1,000 or more employees, including: 261 practitioners, 87 team leads, 183 managers, 165 directors, 34 vice presidents and senior vice presidents, and 315 C-level executives.

About Tray.ai

Tray.ai offers a composable AI integration and automation platform that enterprises use to turn AI and AI agents into standout business performance. The Tray Universal Automation Cloud is a single, AI-ready platform that eliminates the need for disparate tools and technologies to integrate and automate sophisticated internal and external business processes. From prototype to production, with Tray.ai, the development of integrations, the delivery of intelligent apps and the integration of trusted data anywhere is fast, flexible and safe.

Source: Tray.ai