AI Orchestration: Enterprises Juggle Three Platforms

Enterprise AI teams have largely abandoned single orchestration platforms, with the average company now utilizing three. This move is a strategic choice to avoid vendor lock-in and maintain flexibility, but also stems from significant concerns regarding vendor security and permissioning capabilities. Companies are actively seeking to impose their own oversight.

Microsoft leads in current platform usage, but Anthropic's models are highly considered for future adoption. However, challenges persist, particularly concerning token usage and a lack of visibility into agent spending. This ongoing analysis reveals a trend towards hybrid control planes, with over half of respondents expecting them by the end of 2026.

Overall satisfaction with current platforms is high, but ease of implementation and value for money are areas for improvement. Enterprises prioritize security, permissions, and control over agent execution when selecting platforms. Significant investment is being directed towards agent monitoring, debugging, and security enforcement.

Despite advancements, many enterprises still struggle with true multi-step agentic workflows. While control planes are being built, the widespread adoption of autonomous, multi-step agents is still on the horizon for a majority of organizations.