The formula, worked through
Start with customer value: average gross margin per new customer, times the customers a single strong AI recommendation moment could plausibly influence monthly in your category. Then estimate exposure: run ten real buyer questions through the four assistants and count how many produce brand recommendations at all, that percentage is how far the shortlist era has reached your category, and the free plus an hour gives you the number. A B2B firm with $20,000 customers in a category where seven of ten questions already name vendors has a very different justified budget than a local shop where two of ten do. Multiply honestly and most businesses land in one of three tiers below, and the exercise itself, documented, is the business case your CFO actually wants.
The three spend tiers, and who belongs in each
Tier zero, $0 plus founder or marketer hours: right for businesses whose category shows low AI recommendation exposure today, or pre-revenue startups. Run the free baseline and the manual loop from how to start doing AEO and GEO, and re-check exposure quarterly, because it only moves one direction. Tier one, $200 to $800 monthly, platform pricing: right for the broad middle, real category exposure, no spare execution capacity, where software running the loop beats every alternative on cost per moment won; this is where Aethon's published tiers sit, deliberately. Tier two, four to five figures monthly, platform plus agency judgment or dedicated headcount: right only when AI recommendations demonstrably drive contested, high-value categories, and only with the ninety-day evaluation gates from the services guide. The most common budgeting mistake is not overspending or underspending, it is spending tier-two money on tier-one problems because a vendor quoted first and the math never happened.
When to raise, hold, or cut the budget
Revisit quarterly against three signals. Raise when share of tracked moments is climbing and each point of share shows up in branded arrivals and tagged pipeline, per AI traffic analytics, that is a machine converting budget into revenue, and marginal dollars keep earning until share plateaus. Hold when share is climbing but the revenue chain has not confirmed yet, normal in long sales cycles; give it the cycle length before judging. Cut, or re-tier, when two quarters show no share movement despite shipped fixes: something upstream is wrong, usually basket design or category exposure, and more spend amplifies a broken loop. The scoreboard that makes these calls unambiguous is in how to know if your GEO is working, and it is the same one we put in front of our own customers.
