[Ep 044] Agentic AI - Token Economics: Designing Agents That Don’t Blow the Budget

Token stewardship is quickly becoming a make-or-break discipline for Agentic AI because it directly impacts cost, throttling, and whether teams can scale AI usage without losing velocity.In this episode, Greg Banks is joined by Tim Harting and Gavin Franklin to break down token stewardship in practical terms and lay out the core levers every organization can use to control spend while keeping agents effective.What you’ll learnWhere tokens come from in agent systems (prompts, memory, tools, calls, outputs)The 4-tier optimization approach: prompt/context, caching, model routing, and self-hosting tradeoffsWhy token discipline can prevent throttling and protect productivity at scaleWho it’s forDecision makers and technical teams building, deploying, or scaling agentic AI in real operations.Guests + credibilityTim and Gavin Franklin share research-backed guidance and lessons from hands-on internal agent development and usage optimization at Cirrius.CTADownload the episode artifact and use it to start a real cost-and-value conversation inside your organization. (And follow the show for the rest of the Agentic AI series.)Links & ResourcesAgentic AI Playlist of past episodesEpisode ArtifactTim Harting on LinkedInGavin Franklin on LinkedInGreg Banks on LinkedInJason Fowler MusicCirrius Solutions