SOURCE

Single Prompt Content Creation

Properties1
tagslocal-seo, caleb-ulku, course, core-30, ai-content, 8-pass-ai-writing
AUTHOR
Caleb Ulku
ORIGINAL
view original

4 min read

Single Prompt Content Creation

Author: Caleb Ulku | Published: unknown | Source: transcript

Summary

Once the Core 30 gap analysis has named the pages to build, Caleb’s problem is throughput: 20–30 service pages written by hand would take weeks, and each still needs local relevance, keywords, and consistent structure. Most local businesses in a gap analysis, he says, need 15–25 new pages to hit Core 30 coverage. Generic AI copy is not enough; a plumber page that could work in Miami or Denver has failed.

The course asset is a single prompt (two versions: category pages vs. service pages) that is supposed to emit: optimized title tags and meta descriptions, heading structure with local keywords, geographically relevant service descriptions, CTAs, and internal-linking suggestions. Category prompt = broader GBP categories (“water heater services”); service prompt = specific services under those categories (“tankless water heater installation”). He says the prompt also branches by service type: emergency vs. scheduled maintenance, residential vs. commercial, high-ticket (water heater replacement) vs. routine (drain cleaning).

He is explicit that this is foundational content, not final copy — “80% of what you need,” with the last 20% a human pass for tone, local specifics, and business personality. Quality-control rules: regenerate anything that is not city/service-specific; always fact-check technical claims, pricing, and processes. His internal team uses a more complex two-prompt (he also says “two-prong”) process to get closer to publish quality faster; that process, plus the training he wrote for his copywriters, is gated behind Pro.

He closes by reminding students that writing is only the first Core 30 step; deployment with proper internal linking and ongoing optimization is what actually builds topical authority in competitive markets.

Key Claims

  • Typical Core 30 gap is 15–25 new pages; hand-writing them takes weeks.
  • A constrained single prompt can draft local service/category pages at scale with title/meta, headings, geo copy, CTAs, and internal-link suggestions.
  • Use two prompt variants: category pages (GBP categories) vs. service pages (GBP services).
  • Prompt should treat emergency / maintenance / residential / commercial / high-ticket vs. routine services differently.
  • Output is an 80% first draft; human editing is required for tone, local facts, and personality.
  • Reject copy that is interchangeable across cities; regenerate with more specific instructions.
  • Always review AI technical/pricing/process claims against the real business.
  • Agency-internal quality uses a two-prompt process (Pro upsell); this lesson’s single prompt is the student-tier version.
  • Content creation ≠ Core 30 done; internal linking and ongoing optimization still have to happen.

Notable quotes

“The prompt is going to give you 80 of what you need but the final 20 of human touch is what’s going to make the content effective.”

Why citable: Directly contradicts “generate and publish” AI-SEO marketing — including the faster YouTube versions of his own system. Use when aggregating AI-content claims.

“If the content sounds like it could work for a plumber in miami just as well as a plumber in denver have the ai regenerate it with more specific instructions.”

Why citable: His operational test for local relevance, more concrete than “add neighborhoods.”

Connections

Entities mentioned: Caleb Ulku, Google, Claude Concepts referenced: Local SEO, Core 30, GBP–Website Alignment (Consistency Signals), 8-Pass AI Writing Pipeline, Entity-Based SEO

Contradictions / Tensions

  • Single prompt here vs. 8-Pass AI Writing Pipeline and vs. his “train any AI” / multi-pass YouTube workflows. This lesson is the lite student tool; Pro and the public videos sell more passes. Don’t collapse them into one method on the concept page — note the tier split.
  • “Google wants to see” this structure is asserted; the prompt itself is not in the transcript, so we cannot verify the on-page recipe against Google’s own quality rater / helpful-content guidance.
  • Internal-linking “suggestions” from the prompt vs. the Core 30 lesson’s strict GBP-hierarchy linking. If the model suggests nav/footer or unrelated links, that would conflict with his editorial-link doctrine — the transcript does not show the prompt’s linking rules.
  • Commercial: the better process is behind Pro. Weight “this prompt is enough” accordingly.

Notes

Prompt bodies live “below this lesson” in Skool (category + service versions) and are not in the transcript — this page documents the method claims, not the prompt text. Spoken “two-prong” is almost certainly two-prompt. New concept overlap: internal-linking. Does not demonstrate a generated page in this recording.