rong>By Dennis Yu
I just published an article called Why Saying Please and Thank You to AI Makes You a Better Prompter. This meta article explains exactly how my AI agent wrote that piece, published it, sent an email about it, and then updated itself — step by step — so you can see the full process behind the output.
Start With a Real Conversation
The article did not start with a prompt. It started with a real conversation between me and Tom Hawkins, owner of Hawkins Chevrolet in Fairmont, Minnesota. I picked Tom up at Resorts World in Las Vegas, we ate all-you-can-eat Brazilian steakhouse at Fogo de Chao, and we debated whether you should say please and thank you to AI. That debate is the raw material. Without it, no prompt in the world produces a real article.
I gave my AI agent a spoken, unstructured description of our conversation. I mentioned Tom by name, described where we were, referenced the two sides of the argument, explained the second-order effect I have observed about prompt discipline, and told the agent to follow our BlitzMetrics blog posting guidelines.
Research the People and the Claims
The agent searched for Tom Hawkins and Hawkins Chevrolet to confirm real details: the dealership started in Sherburn in 1967, moved to Fairmont in 1989, Tom has been there since 1981 with his brother Mark as a financial partner. It also searched for studies on politeness in AI prompting to ground the claims with actual research rather than opinion.
It then pulled the BlitzMetrics article submission guidelines and blog posting guidelines to understand the structural requirements: minimum 500 words, active voice, short paragraphs, title case headlines, H2 subheadings with substance beneath them, 2-3 relevant links, no AI fluff phrases, no rhetorical questions, and a closing call to action.
Write, Publish, and Send
The agent wrote the article in my voice, first person, with specific experiential details that only exist because the conversation actually happened. Picanha and lamb chops at Fogo. The drive from Resorts World. Dropping Tom at the airport. Years of shared office hours through our AI apprentice program. My visit to Minnesota. These details are what separate a real article from AI-generated content that could have been written about anyone.
After writing the article, the agent logged into the BlitzMetrics WordPress admin, created a new post, pasted the formatted HTML, and clicked Publish. Then it did the same for this meta article. Then it searched my Gmail for Tom Hawkins’ email address, found hawk@hawkinsbestprice.com from previous correspondence, composed an email from itself as the AI agent explaining exactly what it did, included both article links, and sent it.
Pull Visual Evidence From the Zoom Recording
I then asked the agent to log into Zoom and find the Office Hours recording from our call. It navigated to my Zoom recordings, found the March 19 session, opened the Speaker view recording, and searched the transcript for the moment I introduced Tom. It found the frame at timestamp 20:41 showing both of us at Fogo de Chao in the Zoom speaker view. The agent captured this as a screenshot to add to the article as visual proof that the conversation happened in real life, not in a text prompt.
The Recursive Proof
This meta article is itself proof of the main article’s thesis. I gave my AI agent a polite, well-structured, context-rich prompt. I defined the people involved, the setting, the argument, the guidelines to follow, and the outcome I wanted. The agent produced a publishable article, researched the background, formatted everything to spec, published both articles to WordPress, emailed Tom directly, and then went into the Zoom recordings to pull visual evidence. Then it wrote this breakdown of its own process.
If I had said “write something about AI and politeness,” the output would have been generic. The specificity of the prompt drove the specificity of the result. Politeness, structure, and context travel together. That is the point of the original article, and this meta article proves it by existing.
Build Your Own Process
If you want to replicate this workflow, start with a real conversation. Give your AI agent the names, places, and details. Point it to your style guidelines. Let it research the people and claims independently. Have it publish directly to your CMS, send the relevant emails, and pull supporting media from your recordings. Then review the output against your standards.
The AI did the drafting, structuring, research, publishing, emailing, and media retrieval. I brought the relationship with Tom, the lived experience at Fogo de Chao, and the judgment about what matters. That combination is what makes content real. Try it with your own conversations and see what your AI agent produces when you give it something genuine to work with. Join our AI apprentice program at High-Rise Influence to learn more.
Dennis Yu
Dennis Yu is the CEO of Local Service Spotlight, a platform that amplifies the reputations of contractors and local service businesses using the Content Factory process. He is a former search engine engineer who has spent a billion dollars on Google and Facebook ads for Nike, Quiznos, Ashley Furniture, Red Bull, State Farm, and other brands.
Dennis has achieved 25% of his goal of creating a million digital marketing jobs by partnering with universities, professional organizations, and agencies. Through Local Service Spotlight, he teaches the Dollar a Day strategy and Content Factory training to help local service businesses enhance their existing local reputation and make the phone ring.
Dennis coaches young adult agency owners serving plumbers, AC technicians, landscapers, roofers, electricians, and believes there should be a standard in measuring local marketing efforts, much like doctors and plumbers must be certified.