71% of Enterprises Admit Their 'Agents' Are Just Chatbots

A VentureBeat Pulse Research survey of 101 enterprises with 100+ employees reveals AI orchestration is rapidly concentrating on model provider platforms: Anthropic's Claude leads with 40% share (double Microsoft's 18%), followed by OpenAI at 13%. Key drivers are 'model gravity' (21%), meaning native fit with top foundation models; success metrics focus on reliable multi-step execution (task reliability 32%, multi-step workflow management 28%). However, when asked to honestly assess their deployment mix, 71% say only a quarter or fewer of their 'agents' are true multi-step orchestration workflows rather than simple single-prompt chatbot wrappers; only 10% cross the halfway mark. Additionally, over a quarter of enterprises (27%) lack real-time measures to stop runaway agents from draining budgets.