by Michael Stelzner / May 26, 2026
Looking to leverage AI for scaling video content research and production? Curious about building an AI system that generates videos capable of attracting your target audience?
In this piece, you'll learn how to leverage Claude Code to pinpoint videos that are already performing well in your niche, deconstruct the elements behind their success, and produce scripts that reflect your unique voice.
The Rationale Behind the Outlier Video Method
Content creation has always been a demanding process. You film, edit, write scripts, and publish—then repeat the entire cycle. Sandy Lee is all too familiar with this routine. She expanded a language-teaching YouTube channel to 550,000 subscribers across YouTube, Instagram, and TikTok starting in 2018. However, she accomplished this manually, and it nearly exhausted her.
With three children under seven, a full-time job, and client commitments, replicating that success from zero wasn't feasible. So, when she came across Claude Code in late 2025, she didn't merely use AI to craft captions. Instead, she developed a content framework that automates research, pinpoints winning videos, and generates scripts in her own voice.
Within less than a month, she expanded a new YouTube channel from 200 to 11,000 subscribers and generated $10,000 from a client retainer and channel earnings.
The system she built is multi-layered. Sandy describes it as comprising seven AI agents and sub-agents. Think of it as having senior and junior engineers collaborating on your behalf. The senior agents handle overarching decisions, while the sub-agents carry out specific tasks like thumbnail analysis, scriptwriting, and editing guidance.
Below, we concentrate on the three core steps that form the foundation of the entire system: defining your content identity, discovering what's already effective on YouTube, and using AI to convert that research into a script that sounds like you.
The outcome is a content workflow that begins with a daily email featuring outlier videos, proceeds through an AI evaluation of what's working, produces a script in her voice using a proven hook structure, and concludes with Sandy recording content she already feels confident about. The system doesn't supplant her creativity or personal stories. It eliminates the repetitive, time-intensive research and formatting tasks so she can concentrate on what only she can do: appearing on camera.
#1: Develop Your Ideal Customer Profile & Content Pillars With AI
Before you launch Claude Code or any other AI tool, you need to complete one essential preliminary task: creating the brand voice assets that will steer the system.
The Inside Out approach begins with a Japanese concept called Ikigai, which loosely translates to "reason for being." Sandy uses it as a framework for self-discovery.
Take out a notebook and address four questions:
What do you love doing?
What are you skilled at?
What can you get paid for?
What does the world require?
The objective is to pinpoint the intersection where all four answers converge. That intersection is your content identity. It's where your passions meet your skills, what the world needs, and what people are willing to pay for.
A few key points about this step. First, complete it without AI. The aim is to uncover what's genuinely within you, which requires disconnecting. Second, limit yourself to no more than a day. Overanalyzing here is a pitfall. You're not committing to a permanent niche; you're giving AI enough material to work with. Third, your answers will evolve. Sandy adjusted her positioning several times after starting. The key is to get something written down and keep moving forward.
Now, input those notes into your AI tool of choice and ask it to analyze them to produce two items you'll use throughout everything that follows.
Which AI Moves Actually Matter?
The AI Business Society is your reliable guide. Gain expert-led training you can apply immediately. Plus, a community of marketers sharing what's truly effective.
The first is your ideal customer profile (ICP). Based on your notes, AI will identify who is most likely to benefit from your content: their demographics, challenges, search queries, and what they'll pay for. Sandy's ICP turned out to be busy parents and full-time workers seeking financial freedom who believe AI can assist them.
The second output is your content pillars. Your content pillars answer the question you should ask before every video: what type of content does my ideal customer truly need from me? Sandy’s pillars are AI tools and workflows, content creation systems, and the personal brand journey. Every video she produces aligns with one of these pillars.
AI generates the pillars based on the overlap between your Ikigai responses and your ICP.
#2: Research YouTube Content With AI
The next step is to construct a workflow or system that researches YouTube content for you.
Manually, this involves going to YouTube, searching your topic, watching competitor videos, and noting what's gaining traction. Sandy did this herself for years. Now AI handles it for her.
She built a Claude Code system, trained with her ICP and content pillar descriptions, that monitors a curated list of ten YouTube channels in her niche.
3 Days of World-Class Training—Zero Travel!
Couldn't make it to Social Media Marketing World and AI Business World this year? Access all the great content at a fraction of the cost with a Virtual ticket. That includes full access to recordings of every keynote, workshop, and session—the ones people travel thousands of miles to see. Don't wait. Get your Virtual ticket and enjoy actionable content you can watch anytime, anywhere.
Every 48 hours, the system identifies new videos and calculates an outlier score for each recent video across those channels using this formula:
Outlier Score = (Video Views in First 48 Hours ÷ Channel's Average Views in First 48 Hours) × 100
A score significantly above 100 indicates the video is outperforming the channel's typical performance; it's an outlier. This distinction is important. A video might have 500,000 views simply because the creator has a large following; that's not an outlier, it's just scale.
What Sandy is seeking is a video that is dramatically over-performing relative to what that specific channel usually earns. She's looking for a signal that the topic, format, or packaging itself is driving the result, not just the creator's existing audience.
Claude Code connects to YouTube's API to retrieve this data and runs the scoring automatically through an automation platform; Sandy uses n8n. Once the system identifies outlier videos, it emails Sandy a daily digest with the results. Instead of spending an hour each morning scanning channels, she opens her inbox and finds a prioritized list waiting for her.
She reviews the list, selects a video, and proceeds to the next step.
#3: Analyze the Outlier Video With AI
Once Sandy receives her daily email and picks a video she wants to emulate, the AI system takes over again.
Her Claude Code setup analyzes three elements of each outlier video: the thumbnail, the title, and the first 30 seconds of the video, and then describes what's working in each component.
For a thumbnail, it might note that the creator used a side-by-side comparison layout. For the title, it identifies what kind of curiosity is sparked or what promise is made. For the hook, it pinpoints the technique used to capture attention in the initial seconds: curiosity, bold claim, shared pain point, and so on.
Sandy reviews the summary and triggers the next step: script generation.
#4: Craft Your Hook and Script a New Video in Your Voice With AI
Sandy has set up a Claude Code Skill that uses her brand voice information, combined with her content pillars and ICP, to produce a full script in her voice based on the outlier video's message.
The script follows the same topic structure as the outlier video with a similar format and flow, but speaks to her specific audience in her unique way. "Whenever I'm reading the script, I sound more natural," she says, because the script already reflects how she talks and what her audience expects from her.
The 30-second hook is the most crucial part of any video, and Sandy has integrated a specific seven-part hook formula into her Claude Code script prompt. The seven elements, applied sequentially in the first 30 to 60 seconds, are:
Pattern interrupt. Open with something that disrupts the viewer's passive scrolling: a counterintuitive claim, a direct question, or a statement signaling that this video offers something different from what they anticipated. The goal is to stop the click-away before it occurs.
Mirror the viewer. Identify the viewer's pain point in specific, tangible terms. For Sandy's channel, that means describing the actual moment her audience is experiencing: finishing a full workday, coming home, and still having to turn on a camera and deliver something coherent on cue. That level of specificity signals that you understand their daily reality, not just a broad category of frustration.
Opportunity. Clearly state what the viewer stands to gain by continuing to watch. This is the explicit promise you're making: what they'll understand, be able to do, or avoid by staying through the end of the video.
Gap. Establish the stakes without artificial urgency. Sandy's position is straightforward: people who aren't building these skills now will fall behind those who are. That's an accurate observation about where AI adoption is heading, not a sales tactic. State it plainly and let the viewer weigh it for themselves.
Transformation. Describe a concrete before-and-after outcome. Not a vague sense of improvement, but a specific capability the viewer will have by the end of the video that they don't have now.
Authority. Give new viewers a single, well-chosen reason to trust you. Sandy points to her track record: building a channel to 550,000 subscribers, then growing a new one to 11,000 in under a month. One credibility anchor is sufficient.
Transition. Move directly into the first step of the video. No recap, no re-introduction.
Sandy Lee is an AI content strategist and founder of Slee Automation. She helps entrepreneurs use AI to simplify content creation and build scalable personal brands. Her AI Content System community gives members direct access to her strategies, weekly group calls, and accountability support. Follow her on YouTube.
Other Notes From This Episode
Connect with Michael Stelzner @Stelzner on Facebook and @Mike_Stelzner on X.
Watch this interview and other exclusive content from Social Media Examiner on YouTube.
Listen to the Podcast Now
This article is sourced from the AI Explored podcast. Listen or subscribe below.
Where to subscribe: Apple Podcasts | Spotify | YouTube Music | YouTube | Amazon Music | RSS
✋🏽 If you enjoyed this episode of the AI Explored podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.
Stay Up-to-Date: Get New Marketing Articles Delivered to You!
Don't miss out on upcoming social media marketing insights and strategies! Sign up to receive notifications when we publish new articles on Social Media Examiner. Our expertly crafted content will help you stay ahead of the curve and drive results for your business. Click the link below to sign up now and receive our annual report!
Curious About How to Use AI?
Our newest show, AI Explored, might be just what you're looking for. It's for marketers, creators, and entrepreneurs who want to understand how to use AI in their business. It's hosted by Michael Stelzner and explores this exciting new frontier in easy-to-understand terms. Pull up your favorite podcast app and search for AI Explored. Or click the button below for more information.
Tags: AI Explored podcast