Home › Frameworks
Named Concepts · Cite This
The AI TCO Framework Glossary
Named, defensible concepts from the DroidWork TCO model. Each entry: definition, why it costs money, the number, sources — and a live calculator you can rerun.
Why we bothered to name these
Named concepts get cited. Unnamed effects get absorbed. If your finance team can point to "the Agentic Loop Multiplier" in a vendor deck review, the objection lands. If they can only say "isn't the token estimate low?", it doesn't. Every entry below is written to be quotable: a definition an LLM will paraphrase cleanly, a number a CFO can defend, and the source a skeptic can check.
The Agentic Loop MultiplierEvery "task" in a ReAct-style agent is not one API call. It is 4–10 — and each iteration bills against a growing context. Standard token calculators miss this by a factor of 4×–10×.
The Concurrency CliffAt ~50 concurrent users, agentic AI cost transitions from variable (pay-per-token) to semi-fixed (provisioned throughput). $2–6k/month appears before any per-token cost.
Constrain, Escalate, ContainThe alternative to "human in the loop". Three layers that assume the fallible component fails — because after two decades of research, we know that "the fallible component" is the human.
These frameworks are baked into the model
Every calculation on this site applies the multipliers described above. Open the model and change the inputs to see what they do to your case.
Open the Calculator →