APIAI26 D1 1630 SLVL B1 7 PGM P01

Control AI spend at the gateway layer with budgets, anomaly detection, optimization policies, and a shared dashboard for engineering and finance. This session walks through how Kong AI Gateway and Kong Konnect can help teams attribute AI usage, enforce cost controls, detect runaway consumption, and report savings across models, providers, users, and business units. KEY TAKEAWAYS • Budgets can track AI usage without enforcement or actively enforce limits through blocks, notifications, and lifecycle-based rules. • Teams can configure budget actions such as email, Slack, and webhooks to alert or trigger external systems when spend thresholds are crossed. • Anomaly detection helps catch events like rate spikes, baseline deviations, unbudgeted spenders, unattended agent loops, and rising failed transactions before they become major cost overruns. • Optimization policies such as prompt caching and complexity-based routing can reduce token usage and show measurable savings against list-price AI bills. • Attribution depends on mapping AI interactions to a consistent principal through Kong Identity, with metadata and labels used to connect usage to teams, users, agents, and requests. TIMESTAMPS 00:00 - Intro 02:29 - Budget scopes and enforcement 02:57 - Budget actions and notifications 03:47 - Notification channels and webhooks 04:06 - Anomaly detection rules 05:18 - Optimization and savings reporting 06:05 - Cost dashboard overview 08:16 - Data retention question 09:15 - Attribution with Kong Identity 10:46 - Metadata and labels 11:00 - OpenMeter component question 11:17 - Cost management versus billing 12:51 - Konnect platform integration LINKS Explore Kong's AI cost governance solution for FinOps: https://konghq.com/solutions/ai-cost-... Read the guide to AI token cost management: https://konghq.com/blog/enterprise/ai... #AIGateway #AICostManagement #FinOps #KongKonnect #LLM