AOS Scorecard · 7 min

Where is your marketing function on the AI Operator Score?

Sixteen questions across the four AOS levels. Score yourself harder than your board would. At the end, you'll get your level — Curious, Experimenter, Operator, or Architect — plus the three highest-leverage moves to climb one level.

The four AOS levels

AOS 1 — Curious: Marketing is reading about AI. Maybe trying ChatGPT for copy. No documented use.
AOS 2 — Experimenter: Marketing is running AI prompts. Some wins, no system. Brand discipline lives in someone's head.
AOS 3 — Operator: Documented Operator Brief. Prompts run against real data. 1–3 agents shipped on cadence. Measurable outcomes logged.
AOS 4 — Architect: Multi-agent system in production. 5+ agents in cadence. Continuous-context infrastructure. The CMO transition would inherit a working system.

Where is your marketing function on the AI Operator Score?

1. Is your brand voice written down — with specific DO / DON'T rules and forbidden language?

2. Could a smart new hire produce a credible Day-30 LinkedIn post from your brand doc alone?

3. Do you have a documented ICP that an agent could read and apply?

4. Are your buying-committee personas written as system-prompt-ready blocks?

5. How many AI prompts does your team use against real company data on a typical week?

6. Do your prompts pull from a shared context file (Operator Brief or equivalent)?

7. Have you stress-tested marketing copy against simulated ICP personas before publish?

8. How many agent specs (full system prompt + schemas + evals + failure modes) have you written down?

9. How many agents are running on cadence (weekly or more often) in production?

10. Do your agents have hallucination defense built in (verify-before-publish, source-traced outputs)?

11. Do your agents inherit from a single shared context (brand, ICP, KPIs)?

12. Can you defend your marketing ROI to the CFO with the math exposed?

13. Do you proactively watch competitor moves (weekly or better) — or find out when a buyer mentions them?

14. Do you have a written AI marketing policy (governance, autonomy levels, approval gates)?

15. Have you logged measurable outcomes (dollar value) from AI-shipped work?

16. If you left tomorrow, would your successor inherit a working AI marketing system — or rebuild from scratch?

The three highest-leverage moves to climb one level