Essays on how it works
Essays from Heraldic on the decisions behind the product, with the arithmetic shown.
Six essays, published October 30, 2026.
- Worked exampleA 2.0 ROAS at a 40% marginAt a 40% margin, a 2.0 return loses 20 for every 100 spent on ads. The arithmetic, and how agents judge spend by what each sale is worth to you.
- Worked exampleThirty leads a week: when can you call a test?At 30 leads a week, a test of a modest change can’t be called within a month. This essay works through the arithmetic, the rule agents use to call a test, and what they do while they wait.
- WalkthroughAnatomy of a stopped proposalAn agent asked for a bigger budget raise than your rules allow. Code said no, the agent proposed a smaller raise, and every step went on the record. Here it is, end to end.
- How it’s builtWhat it takes to hand an AI a budgetConnecting an AI assistant to your ad and email platforms is useful, and it’s getting better. It still isn’t something you can hand a budget to and walk away from.
- How it’s builtRules belong in codeAn AI asked nicely to stay under budget usually will. “Usually” isn’t good enough for your money or your customers’ inboxes.
- GuideHow to tell if your agents are doing good workAn agent that spends your money should be easy to check. These are the daily signs of whether it’s working.
Read the reasoning, then try the product.
Your work email is all we need to start.