Kong has turned its AI gateway toward the question finance keeps asking. In Take Control of the Economics of AI with Kong AI Gateway, the release starts from a sentence most platform teams will recognise: enterprises “can see what they’re spending on AI, but often can’t explain what they’re spending it on.” A provider bill says how many input and output tokens went to which model. “But businesses don’t operate in tokens. They operate in teams, applications, products, programs, customers, cost centers, people, and increasingly agents.”
The figures in the post are an illustration, not customer data, and it says so: an enterprise that spends 500,000 dollars on AI in a month and knows only the split by provider. The useful claims are architectural. The gateway sits “directly between AI consumers and the models they use,” so it can see the consumer, application, model, provider, tokens and cost on each request as it happens. Kong prices each request individually, because “there isn’t a universal ‘cost per token’”: rates vary by input and output, cached tokens, context-window thresholds and service tiers. Then it attributes the cost up a chain, and the example chain starts with an agent: “Customer Support Agent → Returns & Refunds Workflow → North America → Customer Experience → $27,000.” Budgets, planning and routing to cheaper models follow from having that attribution.
The catalog holds the surfaces this ships on. The Kong provider page lists 139 API pages, and the release lands on the Kong AI Gateway, with the agent traffic it attributes passing through the Kong Agent Gateway and the tools it reaches registered in the Kong MCP Registry. The agentic access profile maps 1,242 operations, 782 of them acting. On the Agent Readiness score, error semantics, reversibility, the event surface, agent skills, delegated identity and protected resource metadata are lit.
The Kin Score is 71.8, exemplar band, carried by access clarity at 84.2 and contract quality at 67.6, and Agent Readiness is 46.9, agent-ready. The MCP server dimension is still unlit, the regression this series reported two weeks ago and filed on our side. Agentic commerce is unlit as well, and this release is the closest Kong has come to it: a per-request price attached to an identified agent is the meter that commerce between agents needs. Idempotency is unlit too, which matters for a cost ledger, because a retried request that is priced twice is a budget error. Kong has made AI spend attributable to the agent that caused it. The contract has to make sure the same request is only charged once.