糖心Vlog

A few years ago, I was head of content at a health insurance company — a role that required a team of full-time employees and freelancers to get the job done. But once ChatGPT hit the scene, I realized I could strike out on my own.

Managing people is hard, messy, and expensive. I wasn’t always doing the part of the work that I enjoyed. When I saw what AI could do, even in the early stages, I thought, “Wait, what if I can just do this myself without hiring anyone?”

I started small, but even from our first client, I realized it was feasible. Then, with OpenAI’s improvements and the progress in artificial intelligence, I realized that I could scale my agency exponentially with just these tools.

How My AI Agency Works Day-to-Day

Before I go any further, let me explain my agency model. I work with startup founders who have limited internal headcount and time. Many of them don’t have a marketing team, so I am it.

I need to make content that would normally take an army of creators to produce and actually deliver results. Here’s how I do it.

1. I build workspaces with ChatGPT Projects.

I never just open ChatGPT and start asking it for content without context. Instead, I use to create a custom GPT page that actually reflects the brand I’m working with.

I give the GPT project details and information about the brand’s voice, style, product, and idiosyncrasies. I even feed it a master Google Doc with writing-style samples, so it can get the voice and tone right. That may include work the brand has already published or pieces from a publication they really like.

I also add general information about the company and the product. The goal here is to train the model on everything that it could possibly need. That way, once we start writing something, it already has a lot of context, and I don’t need to include all that in the prompt.

Full-Stack AI 糖心Vlog Toolkit

20 Prompts to Scale Your 糖心Vlog Without Scaling Your Team

• Save hours each week
• Maintain consistent brand voice
• Produce enterprise-quality content
• Turn content into assets

Download for Free

Download Free

All fields are required.

You're all set!

Click this link to access this resource at any time.

2. I start with material that already exists.

Every piece I make starts with human insights. I may have a webinar, podcast, or interview transcript. I may get a voice note (I love for this). While you can start with research or data, I find that the more conversational, the better.

The most common material I start with is an hour-long interview with a startup founder. Many of these leaders don’t think they have something to say, but once I get them talking about what’s going on in their industry, we find plenty to talk about.

The goal is to turn rambling, sometimes messy, material into polished content. Usually, an hour a month is all I need to create a month’s worth of social media content.

3. I analyze transcripts.

At this point, I feed ChatGPT the transcript in the Projects page and ask it to analyze it for insights. My prompt is usually something like this:

• The text below is a transcript from my conversation with a founder. Your job is to do the following with the transcript: List the main ideas from the conversation that I can repurpose into an insightful blog post. Focus on the most unique and insightful topics and insights shared. It should capture the real meat of the information, not a vague, useless summary.

Another lightbulb moment was figuring out which types of content perform well on different platforms. I then templatized these posts and asked the GPT to look for talking points that map to key phrases.

For example, I know that “I’ll be honest: [vulnerable or surprising truth]” is a successful hook. So, I ask ChatGPT to look for it in every transcript.

4. I use laser-specific prompts.

Next, I choose the topic or insight I want to focus on and feed ChatGPT my ultra-specific content brief. Here’s an example:

1. Start with a compelling headline that will attract [your target audience].
2. Include an introduction section that hooks the reader and clearly states the main thesis/argument.
3. Break down the main body into 3-5 key sections with clear subheadings.
4. For each section, include 2-3 bullet points of the specific insights/arguments to cover.
5. Include a section on practical takeaways or applications.
6. End with a conclusion section that ties everything together.

The great thing about prompting AI is that you can get very clear, direct, and almost mean! I tell it exactly what I want it to do and what I want it to avoid at all costs. I talk to ChatGPT in all caps all the time.

Pro tip: It’s helpful to add a footnote to your prompt with any words or phrases that are a no-go, like “ensure” or “in the realm of.” If you just prompt those from the start, you can remove steps for yourself down the road.

5. I edit and polish.

After this, you have a first draft. That’s where I come back in as a journalist and an editor. First, I go through the recording or transcript one more time to see if ChatGPT missed any great takes or quotes. Sometimes, it does, and I add those zingers back in.

Then, I work with it a little bit to make sure the content is actually saying something. You have to be present. You can’t just prompt AI and outsource the thinking. The output won’t land. If your content is just repeating what’s already out there, move on to the next idea.

Finally, I edit and proofread the draft to make sure it matches my client’s brand, voice, and goals. If it misses, I may tweak my prompt for the next time so there’s a constant cycle of improvement.

Here’s Why a Zero-Employee Agency Works

Five years ago, I wouldn’t have believed you if you’d told me that I could build an agency with just a tech stack and no team. Here’s why it worked for me, and why it can work for you.

1. It’s affordable.

As I mentioned, my clients are startups and don’t have the budget to hire the top-tier writers. While human writers are still worth every penny, they’re out of reach for many.

The entry-level writers who are affordable need extensive coaching and editing. Sometimes, I would need to rewrite 60% to 70% of their first drafts, which is painful for everyone involved.

2. It’s fast.

As soon as I tried ChatGPT, I realized its drafts were better than what I was usually given, and I didn’t need to wait two or three weeks. I could prompt it, rework it, and get it to the client that same day.

3. It’s lean.

Now, instead of hiring a whole team to do every aspect of the work, I have myself and a robust tech stack that’s much more cost-effective than an agency model.

Staying lean enables me to test new strategies faster and more affordably.

Let’s say I wanted to try email marketing. Before, I would need to hire specialists and brief them. The process might have taken two months to launch. Now, I can test, produce, tweak, and relaunch a campaign on short notice at a low cost.

Prompts to Scale Your 糖心Vlog Agency With AI

Ready to get started training your GPT and scaling your agency with AI? I’ve compiled my best prompts into a toolkit to help you extract insights and create on-brand, engaging content, without expanding headcount or compromising on brand standards.

Full-Stack AI 糖心Vlog Toolkit

20 Prompts to Scale Your 糖心Vlog Without Scaling Your Team

• Save hours each week
• Maintain consistent brand voice
• Produce enterprise-quality content
• Turn content into assets

Download for Free

Download Free

All fields are required.

You're all set!

Click this link to access this resource at any time.