Enterprise AI Agents: Orchestration Platforms Chosen, But True Automation Lags
A new VentureBeat Pulse Research report reveals that 101 enterprises are consolidating their AI agent orchestration onto model-provider platforms, with Anthropic's Claude leading significantly at 40%. Microsoft follows at 18%, and OpenAI at 13%. This selection is primarily driven by "model gravity," meaning enterprises choose platforms that align with their preferred state-of-the-art base models. Success is measured by reliable multi-step execution, with task completion reliability (32%) and multi-step workflow management (28%) being key metrics.
However, a substantial gap exists between the ambition for agent orchestration and its current reality. A significant 71% of enterprises report that a quarter or fewer of their deployed "agents" are true multi-step orchestrated workflows, with most still functioning as simple chatbot wrappers. Looking ahead, 51% of organizations anticipate a hybrid control plane by 2026, combining provider-native services with external orchestration, largely to avoid vendor lock-in, which is the primary fear for 35% of respondents. Investment is currently focused on agent workflow tooling (34%) and security/permissions (25%).
A critical concern is the lack of real-time fiscal control, with over a quarter (27%) unable to prevent runaway AI agent costs before they incur. While enterprises rate their current orchestration platforms at an average of 3.94 out of 5, indicating provisional acceptance, 96% plan to change their approach within the year, suggesting dissatisfaction and a search for more robust solutions. The current focus on model providers for orchestration, rather than open frameworks, highlights a trend towards integrated solutions, but the maturity of actual multi-step automation remains a significant challenge.
The research highlights a common pattern in emerging technology adoption: the rapid consolidation of foundational infrastructure onto dominant platform providers, driven by the perceived quality of the core technology (model gravity). However, this consolidation is outpacing the development of sophisticated application layers, leading to a "chatbot trap" where advanced orchestration capabilities are desired but not yet fully realized in deployed "agents." Enterprises are prioritizing vendor flexibility through hybrid control planes to mitigate future lock-in, indicating a cautious approach to long-term AI strategy. The significant portion of organizations lacking real-time cost controls points to a critical governance gap, suggesting that the focus on deployment speed and model capability has overshadowed essential operational and financial management. This dynamic sets the stage for future innovation in both agentic workflow design and the development of robust, cost-aware AI governance frameworks over the next decade.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.