TLDR: The Connected Enterprise (CCE) 2025 conference highlighted a significant gap between the ambitious vision for AI agents and their current practical implementation in enterprises. While autonomous AI agents are seen as the next frontier, businesses are grappling with platform fatigue, a lack of strategic clarity, and the complex realities of deploying these systems securely and effectively. Experts at the conference emphasized the need for a more grounded approach, focusing on managing AI agents like human teams and demanding tangible value from technology vendors.
The Connected Enterprise (CCE) 2025 conference, hosted by Constellation Research, brought to light a critical divergence between the aspirational ‘dream’ of AI agents and their current ‘reality’ within the enterprise landscape. A prevailing sentiment among CxOs was that while the potential of AI agents is immense, widespread, fully autonomous deployment remains a year or more away. This realization is prompting a re-evaluation of strategies and a growing weariness among IT buyers.
One of the most prominent themes at CCE 2025 was ‘platform fatigue.’ Enterprise leaders are increasingly overwhelmed by the proliferation of AI platforms and the constant pressure to adopt new solutions. Many expressed a desire to develop internal systems that could reduce reliance on costly SaaS vendors. Fiona Tan, CTO of Wayfair, articulated this sentiment, stating, ‘We’re looking for a horizontal platform partner that we can work with.’ This was echoed by analyst Liz Miller, who observed, ‘Every time someone would be like, here’s what you gotta do with AI. You’ve got to pick a platform and go. And then everyone was like, okay. What platform?… but I’m really sick of platforms.’ This exhaustion is driving a fundamental re-evaluation of IT strategies, including the perennial ‘build vs. buy’ dilemma.
Despite the current challenges, the long-term vision for AI agents remains compelling. These autonomous systems, capable of making decisions without direct human intervention, are viewed as a crucial step beyond current AI models, moving towards more business-relevant applications rather than the broader, often hyped, concept of Artificial General Intelligence (AGI). However, the immediate reality is that enterprises are encountering significant hurdles. Multi-step agents designed to solve complex, long-term problems are still largely elusive, and many early AI agent products fail to deliver measurable productivity gains, leading to user disengagement.
Experts at CCE 2025 stressed that AI agents would require careful management, akin to human teams. ‘AI agents need a manager,’ and ‘AI agents will team with humans,’ were key takeaways. Presser, a speaker at the event, humorously but pointedly noted, ‘Agents are for work. Humans are for love.’ This highlights the necessity of selecting appropriate tasks for agents and ensuring they complement each other within a broader operational framework. The ultimate goal is to integrate agentic AI as a ‘circulatory system for processes and business operations,’ enabling seamless business flows driven by culture, roadmaps, customer engagement, and ecosystems.
However, the complexity of building and deploying robust AI agents in an enterprise setting cannot be underestimated. Challenges include ensuring security, establishing comprehensive observability, and developing effective orchestration layers for deployment. Many current ‘AI agents’ are, in reality, more deterministic workflows or advanced AI assistants, rather than the sophisticated, fully autonomous entities envisioned. This gap between expectation and current capability suggests that while there’s considerable upside, navigating the evolving landscape requires a pragmatic approach.
Constellation Research analyst Martin Schneider underscored the urgency for organizations to prepare for these interconnected, ‘agentic’ ecosystems. He stated, ‘The AI exponentials are here… services will be provided… pricing models are changing,’ and emphasized, ‘What your agentic AI orchestrator is… because these agents are here. They’re getting used to it, and they are talking to each other, and it’s multi-platform.’ This indicates a fundamental shift in how digital operations will be structured, demanding immediate attention from businesses.
Also Read:
- Navigating the AI Landscape: Why Businesses Must Diversify Beyond Single Models and Avoid Over-Reliance
- Paradox in AI Job Market: Specialized Agent Developer Roles Scarce Amidst Explosive Growth
Esteban Kolsky, Chief Distiller at Constellation Research, also weighed in on the role of AI in decision intelligence for executives. He argued that simply querying AI is insufficient; a deep understanding of domain and language correlation is essential. Kolsky suggested that analyst firms, by steering away from this hands-on process, have faced an ‘unfair knee-jerk reaction’ in the market, but acknowledged that their past approaches contributed to this perception. With only about 8% of businesses in Europe and 5% in the U.S. currently leveraging AI, there’s a clear need for more effective integration and a focus on tangible returns on investment.
