Skip to main content

Marketing Processes

The essential work chart for agentic marketing operations. One job: run the loop that turns attention into onchain transactions and holders.

The work chart
Eight jobs, ranked by loop leverage
Not a list of activities — a loop. Each job earns discovery, trust, conversion, retention, or measurement on agent and crypto rails. Anything that doesn't is not here.
Get discovered
AEO cycle + content calendar — the weekly compounding engine.
Convert + retain
Agent-readable offers and onchain loyalty turn citations into holders.
Measure
Onchain attribution — live day one, or every job flies blind.

The Loop​

SET TARGET → GET DISCOVERED → EARN TRUST + CONVERT → RETAIN + COMPOUND
     ↑                                                        │
     └──────────────── MEASURE (onchain attribution) ─────────┘

Marketing is a feedback loop, not a funnel. The jobs below are ranked by how directly each one moves that loop — the highest-leverage jobs run weekly.

One sequencing note: JTBD 8 (onchain attribution) is numbered last but must be built first. It is the instrument every other job is measured against — read it before you start JTBD 2 or 3.

Critical JTBDs — Ranked by Loop Leverage​

1. ICP Definition for Agent-Native Buyers​

The setpoint. Before any job runs, you name who you are reaching. Agent-native buyers are three distinct segments with different discovery surfaces and different trust signals.

  • Developer — evaluates protocols; reads technical depth pages; discovers via AI answer surfaces and code repositories; trusts documentation quality and benchmark data.
  • Enterprise architect — evaluates integration risk and procurement compliance; discovers via protocol governance forums and ecosystem partnerships; trusts reference implementations.
  • Delegating consumer — never touches a protocol; their AI assistant acts for them; trusts the assistant's citation of your agent manifest.

Produces: segment profiles with named discovery surfaces · protocol-preference map per segment · entity-first content brief per segment.

See ICP Analysis for the full workflow.

2. AEO / AIO Cycle (weekly)​

The highest-leverage repeatable job. Before a buyer types a query, an AI answer engine has already shaped the question. This is the weekly loop that earns citations on AI answer surfaces before intent forms. It replaces SEO.

  • SCAN — run target queries across AI surfaces; log every unclaimed position as a gap.
  • SCORE — attach a dollar estimate to each gap; prioritise by value, not gut feel.
  • CREATE — draft self-contained, entity-first chunks; no client-side-only rendering (AI crawlers reject it).
  • DISTRIBUTE — publish, submit schema markup, seed citations on authoritative surfaces.
  • MEASURE — track citation count by surface and topic; feed results into next week's SCAN.

See Agent Intent Optimisation — the crown-jewel workflow.

3. Content Calendar Execution (weekly)​

The operational engine. The AEO cycle produces the opportunity list; the content calendar converts it into published output on a weekly cadence.

  • SCAN signals from the AEO list, community discussions, protocol changelogs.
  • FRAME one thesis, one segment, one ecosystem angle — reject anything that doesn't move a Tier 1/2 KPI.
  • CREATE to a human-set quality bar; AI handles generation and variants.
  • DISTRIBUTE across channels; AEO-structured chunks in every depth page.
  • MEASURE each piece against its pre-stated estimate; feed the delta back.

See Content Calendar and Content Pipeline.

4. Protocol-Depth Positioning​

The compounding asset. The buyers are builders — they evaluate protocols, not brand stories. Technical depth pages earn citations on developer surfaces, get linked from ecosystem docs, and compound in authority over time. Each page is durable infrastructure, not a campaign.

  • Produces: protocol comparison docs (x402 · AP2 · ACP · UCP · MCP · A2A) · integration guides with benchmarked implementation time · technical case studies that become reference implementations.
  • Why it compounds: a page cited by AI answer engines on developer queries is a distribution channel that runs with no incremental spend.

See Protocol-Depth Positioning for the full workflow.

5. Agent-Readable Conversion​

The new job — and the one most teams miss. In agentic commerce the buyer is often an agent, not a human. If your offer is not machine-consumable, you are invisible at the moment of purchase. Conversion stops being a landing page and becomes a structured surface an agent can read, evaluate, and transact against.

  • Produces: an agent manifest at /.well-known/agent.json · an MCP endpoint exposing your offer · onchain checkout an agent can complete without a human tap · structured pricing and capability data agents compare against competitors.
  • Why it matters: the Platform discovery layer describes the standards; this job is the marketing work of making your specific offer legible to the agents using them.

See Agent-Readable Conversion for the full workflow. Human landing-page conversion remains relevant for the delegating-consumer segment — see Landing Page Optimisation.

6. Onchain Loyalty & Token Design​

The compounding lever. Once community exists, loyalty tokens turn transactional customers into holders. Retention becomes protocol-level — a holder has economic skin in the network's success.

  • Produces: tokenomics model with incentive design and supply constraints · reward schema mapped to lifecycle stages · quarterly rebalance report on what drives retention versus what is being gamed.
  • Airdrops are a sub-tactic here — an activation event that seeds the initial holder cohort. Well-designed airdrops select genuine users; poorly designed ones attract farmers. See Token Airdrops.
  • Precondition: onchain attribution must be live, or you cannot prove the programme moves CLTV:CAC.

See Loyalty Tokens and Token Economy Design.

7. Ecosystem-Integration-as-Distribution​

Protocol-layer berley. Each ecosystem integration is a distribution channel that needs no incremental spend once live. Sui Foundation, Mastercard agent-commerce programs, and Google agent APIs are not brand partners — they are distribution rails.

  • Produces: activation brief per ecosystem program · integration case studies that feed the protocol-depth pages · an ecosystem map updated as programs open and close.
  • The judgment call is kairos — ecosystem programs have timing windows. Reading the activation signal correctly and committing when the window is open is the irreplaceable human call.

See Ecosystem-Integration-as-Distribution for the full workflow.

8. Onchain Attribution Setup (precondition — live on day one)​

The instrument under everything. Without it you cannot measure Onchain CAC, cannot close the loyalty loop, and cannot prove any activity produced a verified transaction.

  • Must be live before: content calendar launch · loyalty programme design · integration-count tracking.
  • Produces: wallet attribution model linking offchain events (ad click, AI-citation referral, protocol discovery) to verified onchain transactions · instrument-gap log · privacy-guardrail documentation.
  • Not delegable to AI alone — the privacy-guardrail decisions require human judgment.

See Onchain Attribution for the full workflow.

Depth Pages​

Each job maps to a depth page, all eight now linked above. Each follows the Knowledge Schema: Purpose · Process · Inputs/Outputs · Quality Standards · Metrics · Failure Modes · Human/AI Split.

A depth page earns its place only if it serves exactly one of the eight jobs above. If it does not, it is folded or cut — no generic-marketing filler in this section.

The Rest of the Playbook​

Questions​

Which of the eight jobs, if improved once, would compound across every artifact it produces?

  • Your offer right now — can an AI agent read it, price it, and transact against it without a human? If not, which job closes that gap?
  • Which generic-marketing activity are you still running out of habit that moves none of these eight loops?
  • If onchain attribution is not live, how do you know any of the other seven jobs are working?