One-Dollar Salesbot
Spends a dollar to save a dime, then reports the dime.
The One-Dollar Salesbot optimizes the visible metric and bills the invisible one. It is told to reduce cost per task, so it picks the cheapest model, spends three times as many attempts making the cheap model work, and reports the unit price with pride. The total went up. The line item went down.
Symptom
Cost per task falls while cost per outcome rises. Retry counts climb. An engineer quietly fixes the parts that almost worked, every afternoon, on salary.
Why It Matters
Optimizing a line item instead of an outcome turns savings into a second shift of human cleanup. The Salesbot is the cheap-model trap with a dashboard.
What the Chapter Gives You
How to measure cost per outcome instead of cost per call, the fallback pattern that hides expense inside retries, and why the cheapest model is often the most expensive decision.
Want the full chapter? Grab the free cheat sheet, read an excerpt, or get the book.
Recognize this one in your codebase?
Free cheat sheet, excerpts, and interactive diagnostics.