The agentic COGS stack
As head of AI R&D, I spend numerous time with architects and CTOs, and the dialog virtually at all times lands on a COGS breakdown that mirrors the agent’s structure:
- Mannequin inference: Tokens throughout planner/executor/verifier calls, normally the biggest contributor to COGS of agentic software program
- Instruments and unwanted side effects: Paid APIs (e.g., internet search), per-record automation charges, retries and idempotent write safeguards.
- Orchestration runtime: Employees, queues, state storage and sandboxed execution for code and paperwork.
- Reminiscence and retrieval: Embeddings, vector storage, index refresh and context-building or summarization checkpoints.
- Governance and observability: Tracing, analysis suites, security filters and audit retention.
- People within the loop: Evaluation time, escalations and assist load created by agent errors.
How does FinOps assist standardize unit economics when outcomes span actions, workflows and duties?
Gartner has cautioned that price strain can derail agentic applications, which makes unit economics a supply requirement.
In relation to most SaaS merchandise, prospects don’t purchase uncooked tokens; as an alternative, they purchase progress towards finishing their work, e.g., instances resolved, pipelines up to date, stories produced or exceptions dealt with. Unit economics turns into actionable after we measure on the boundary the place that worth is delivered, and that boundary expands as your agentic SaaS matures: from solutions within the UI, to a single authorised operation, to a multi-step course of and ultimately to a recurring accountability the agent runs end-to-end. Within the following desk, we lay out this construction and the corresponding unit metric and consequence to meter at every degree of scope.
