Geordie launches Cost Intelligence to tie AI spending to agent activity
SiliconANGLE Duncan Riley
Geordie added a tool that shows which AI agents are running up the bill. That turns AI spend from a blob of tokens into something a company can actually pin on a workflow.
Based on reporting by SiliconANGLE, Duncan Riley — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Geordie AI Ltd. has pushed its agent security product into a new job: cost control. Its new Cost Intelligence feature ties AI spending back to the agents and workflows that created it, instead of leaving finance teams staring at token totals and model bills with no clear culprit.
That matters because AI costs can hide in plain sight. Geordie says one retailer was burning $1 million a month by sending work to models that were pricier than the task needed, while a bank saw an agent trapped in a tool loop consume 50% of its monthly AI budget in a single day. Those are the kind of numbers that make “usage” look a lot less harmless.
The company says it can do this because its platform already sits close to the agent to monitor behavior for security and governance. That same position lets it see what the agent did, how many tokens it used and which models it called. It can then trace a cost line back to the agent, the task and the person who owns it. Measurement happens across cloud, code and endpoint, which Geordie says gives a broader view than tools that only watch traffic through one gateway or router.
The release bundles five capabilities. Spending from connected platforms is pulled into one view, and teams can drill from company-wide totals down to platform, team, user, model, individual agent or workflow. It also tracks caching efficiency per agent, maps token use against spend so busy work doesn’t get mislabeled as waste, and looks for anomalies like infinite loops, oversized models and token-heavy sessions that produce little value.
Geordie co-founder and chief executive Henry Comfort says tokens and invoices only show consumption, not the work that drove it. Customer Wood Partners LLC says that visibility changed the internal question from how much is being spent to whether the spend is worth it. That’s the right direction. AI budgets have spent long enough hiding behind mushy dashboards and vague “adoption” talk; sooner or later someone has to ask which agents are actually earning their keep.
My take — AI-written commentary, not fact-checked reporting
This is the more honest version of AI governance: follow the money, not the slide deck. Security vendors love to sell fear, but the real pain for many companies is finding out their shiny agent is a small, expensive chaos machine. If a tool can expose that without pretending every busy model is doing heroic work, it’s probably useful for once.
Read more about this at: SiliconANGLE
