For the past two years, the enterprise AI race has largely been a competition between models: OpenAI's GPT, Anthropic's Claude, and Google's Gemini. But a strategic shift is underway. The next battleground is not about which model answers prompts best, but who controls the layer where AI agents plan, call tools, access data, run workflows, and provide security auditability. New survey data from VB Pulse reveals early market dynamics in this emerging category. Here are key questions and answers to understand this transition.
1. What is the new battleground in enterprise AI, and why does it matter?
The new battleground is the agent control plane—the infrastructure layer that orchestrates how AI agents operate within enterprise environments. Unlike earlier focus on model performance, enterprises now care about where agents execute tasks, integrate with existing systems, and maintain governance. This matters because agents are becoming the operational machinery of AI work; decisions made here affect security, scalability, and auditability. Companies like Microsoft, OpenAI, and Anthropic are racing to dominate this layer, recognizing that controlling orchestration means controlling enterprise AI adoption at scale.
2. What do the latest VB Pulse survey data reveal about enterprise agent orchestration adoption?
VB Pulse's independent Enterprise Agentic Orchestration tracker, surveying qualified technical decision-makers, shows Microsoft leading with 38.6% primary-platform adoption in February 2025 (up from 35.7% in January), driven by Copilot Studio and Azure AI Studio. OpenAI holds second place at 25.7% (up from 23.2%) with Assistants and Responses API. Anthropic entered the tracker for the first time at 5.7% for tool use and workflows, up from 0% in January. While Microsoft and OpenAI dominate, Anthropic's emergence signals growing interest in native orchestration beyond the model layer.
3. Why is Anthropic's 5.7% appearance considered strategically significant despite its small size?
Though a jump from 0% to 5.7% among 70 respondents represents only four enterprise users, the strategic significance lies in the shift from model to orchestration. It marks the first measurable evidence that Claude is moving from being a pure model into a managed runtime for enterprise agents. Early footholds often set the stage for broader adoption, especially as enterprises seek alternatives to Microsoft and OpenAI. As Tom Findling, CEO of Conifers, notes, “Models and agent frameworks have matured enough together that enterprises are shifting focus beyond model quality to the control plane around it.”
4. How does the quote from Tom Findling summarize the enterprise AI convergence?
Tom Findling highlights that enterprise AI is at a convergence moment. He states, “In security operations, we’re seeing the competitive advantage move toward platforms that can orchestrate agents, leverage enterprise context, and provide governance and auditability across customer environments.” This aligns with the data: enterprises now value orchestration capabilities—like planning, tool access, and data integration—over raw model quality. The control plane becomes the decisive factor for security, compliance, and operational efficiency, driving adoption of platforms like Microsoft's, OpenAI's, or emerging ones like Anthropic's.
5. What are the implications for enterprises choosing between AI platforms today?
Enterprises must evaluate not just model performance but the entire orchestration stack. Microsoft offers deep integration with existing enterprise ecosystems (Azure, Copilot). OpenAI provides strong API infrastructure for building custom agents. Anthropic is positioning its own managed runtime. The choice affects future flexibility, security audits, and vendor lock-in. Current leaders have early advantages—Microsoft with distribution, OpenAI with installed base—but small entrants like Anthropic indicate the market is still fluid. Decision-makers should prioritize platforms that offer governance, auditability, and hybrid interoperability.
6. Are models still important, or has the focus completely shifted to orchestration?
Models remain fundamental—they power reasoning and generation—but orchestration is now the competitive differentiator. As the VB Pulse data shows, enterprises are selecting platforms based on how well agents plan, access data, and comply with security policies. The model war is still alive, but it’s no longer the sole factor. A superior model on a weak orchestration platform loses to an adequate model on a robust control plane. The future belongs to providers who can combine strong models with seamless, secure agent orchestration.
To explore further, see current market leaders or the convergence viewpoint.