Cisco ships Splunk Agent Observability with real-time token tracking

Cisco Systems unveiled expanded Splunk Agent Observability on September 15, 2026, introducing real-time visibility into AI agent performance, token expenditure, and coding-agent adoption across enterprises Cisco Newsroom.

The centerpiece is Tokenomics, a new solution within Agent Observability that tracks and attributes token spend across AI agents and employees' use of coding agents such as Claude Code, Codex, and Cursor Cisco Newsroom. The feature ties adoption metrics directly to business value, giving organizations granular cost visibility as agent deployments scale across their infrastructure.

Full-Stack AI Monitoring

Splunk Agent Observability monitors the entire AI stack—GPUs, vector databases, memory consumption, and orchestration frameworks—to evaluate agent behavior across all layers. This end-to-end visibility addresses a critical gap for enterprises running multiple coding and autonomous agents: the inability to correlate agent actions with resource consumption and financial impact. Organizations can now track which agents are consuming the most tokens, how memory allocations affect performance, and which orchestration patterns optimize both cost and latency.

The platform is now available in both Splunk Observability Cloud and Cisco Cloud Control Cisco Newsroom, enabling on-premises and cloud-hosted deployments. This dual-track release allows organizations to integrate agent observability into their existing Splunk infrastructure without forced migration.

Market Context

The launch arrives as enterprises scale agent deployments in production but lack standardized tools to govern and cost-allocate agent workloads. Tokenomics directly addresses this operational blind spot by providing the first real-time attribution layer for AI agent spending—a critical input for chargeback models, budget forecasting, and ROI calculation in large organizations managing dozens or hundreds of agents across teams.

By extending visibility to specific agents like Claude Code and Cursor—naming them explicitly in the announcement—Cisco signals explicit support for major third-party coding agents, not just proprietary tools. This interoperability stance positions Splunk Agent Observability as agent-agnostic infrastructure rather than a closed platform, a strategic move as the agent ecosystem diversifies.