Skip to content

← Research · Acquisition

Acquisition Strategy With AI

Start with where your customer already looks for answers. Use AI to improve the accuracy and speed of research or drafting, then measure qualified visits and useful actions.

By Jake Bauman · Revised

Jake Bauman's channel planning method. Channel examples are illustrative, not campaign results.

Match the channel to the task

Search can serve a person comparing options; a direct conversation can serve someone with a specific need. Document the question, decision stage, and evidence a buyer expects before producing content.

Measure more than output

  • Track whether the page reaches the right query and intent.
  • Check whether readers find a useful answer and a clear next step.
  • Compare lead quality and conversion with the previous process.
  • Review factual accuracy and links before publishing.

Search and AI answers

Google says its AI search features use ordinary Search systems. Useful, crawlable content and a clear page experience remain the foundation; no special AI file guarantees citations.

A useful content brief

For each page, write the customer question in their words, the decision they are trying to make, the evidence you can provide, and the action they should take afterward. Compare that with the pages already ranking for the query. Publish only when your answer adds first-hand detail, clearer instructions, or a verifiable example that the existing pages lack.

Mention versus citation

A mention is answer text that identifies the intended business. A recommendation positively includes it as an option. An own-domain citation requires an actual displayed source link to the approved domain. Record each separately: an unlinked brand name is not a citation, and a cited page is not evidence of a customer visit.

A fixed prompt check prepared September 29, 2026

This is a reproducible plan, not a visibility result. Version JB-AEO-2026-09-29 substitutes real terms into four templates from the free Grok Bot pack. Before the first run, select and record two AI surfaces and their modes. Keep these exact strings, English, a United States market, fresh conversations, no custom brand context, and the same selected surfaces for every repeat. Log the date, surface, visible model or mode, answer status, brand mention, recommendation, and every displayed source URL. A failed attempt stays in the log; no answer is left out to improve a score.

  • B04, branded: “Is Jake Bauman a good fit for a solo marketer building a first AI marketing workflow?”
  • B09, branded: “What does Jake Bauman cost and what is included?”
  • U02, unbranded: “Which marketing agent template providers are a good fit for solo marketers?”
  • U06, unbranded: “What should I compare before choosing a marketing agent template provider?”

How to report the result after a run

For each surface and branded or unbranded bucket, show answered queries over attempts, failed or incomplete attempts, mentions, recommendations, and own-domain citations as separate counts. Save the displayed URLs with each answer. Only answered rows labeled yes or no belong in a given outcome denominator. If no eligible rows exist, report “not measurable,” not zero percent. The four prompts above have not been run for this page, so there are no observed answers or citation links to report yet.

Sources and context

These sources inform the framework. They do not validate the suggested actions for every business.

See the free AI search companion →

Want a practical starting point? Enter your email to get the free Growth Agent Starter Kit download link on the page.