Geordie adds cost intelligence for AI agents

Geordie on September 9, 2026 launched Cost Intelligence, a new capability that ties AI spend to the agents, activities, and workflows generating it. The feature is designed to help enterprises spot waste, measure agent efficiency, and manage financial risk alongside security risk.

Why it matters: - AI spending is often tracked by tokens and invoices, but those numbers do not show which agent or workflow created the cost. - Geordie’s new capability gives AI Ops teams a clearer view of financial risk, waste, and agent-level efficiency across the enterprise. - The feature is meant to support broader agent risk management, not just security oversight.

What happened: - Geordie announced Cost Intelligence on September 9, 2026, in New York. - The capability connects agent usage and activity directly to the costs they generate. - Cost Intelligence is generally available starting today. - More information is available in Geordie’s announcement.

The details: - Geordie’s architecture sits close to the agent, which gives the platform visibility into behavior, token consumption, and model usage. - AI Ops teams can trace a spend change back to the agent that acted, what the agent did, and who was responsible. - Geordie measures spend at the individual-agent level across cloud, code, and endpoint. - The platform aggregates AI spend across connected platforms instead of only tracking activity routed through a single gateway or model router. - Key capabilities include cross-platform spend aggregation, multi-dimensional breakdowns, cache-utilization insight, token-efficiency mapping, and cost anomaly detection. - Multi-dimensional breakdowns let teams drill from an organization-wide total to platform, team, user, model, individual agent, or workflow. - Cache-utilization insight tracks per-agent caching efficiency and flags under-caching agents. - Token-efficiency mapping links spend and volume per agent so useful high-volume agents are not mistaken for wasteful ones. - Cost anomaly detection identifies agents stuck in infinite loops, tasks routed to oversized models, and sessions burning tokens without much productive output. - Wood Partners VP of Technology Jacob Sweat said Geordie improved visibility into AI spend and changed the discussion from how much is being spent to whether the spend is worth it. - Geordie co-founder and CEO Henry Comfort said tokens and invoices do not show what work drove the consumption, and that Cost Intelligence is designed to connect spend to individual agents and their activity. - Geordie co-founder and chief AI officer Hanah-Marie Darley said the risk from autonomous agents includes operational, financial, and accountability risk.

Between the lines: - Geordie is positioning agent monitoring as a financial-control problem as much as a security problem. - The examples of model-routing waste and runaway tool loops point to a practical issue: AI costs can spike even when usage looks normal at a high level. - The product’s value depends on instrumentation that can follow agent behavior across the AI estate, not just at the point where costs are billed.

What's next: - Enterprises using Geordie can now evaluate which agents create value and which ones create avoidable cost. - The company is likely to keep expanding its agent-risk platform as organizations push more autonomous work into production. - Geordie says Security and AI teams use the platform to discover agents, map connections, understand behavior, and steer agents away from risky actions. - Geordie says it is relied on by Fortune 500 companies and other large enterprises, and it won the 2026 RSAC Innovation Sandbox.

The bottom line: - Geordie is trying to make AI spend legible at the agent level, so enterprises can manage cost, accountability, and risk in one place.