Environments for building, running, and orchestrating AI agents that execute multi-step GTM work. They manage memory, tool access, and task routing so an agent can plan and act across connected systems. In an AI-native GTM system, this layer coordinates the agents that operate the pipeline day to day.