How to Use AI in Your Business Without Losing Your Edge
May 11, 2026
I built a 7-figure business working 8-10 hours a week. AI didn't replace me. It freed me to do the work only I can do.
When I first started experimenting with AI, I treated it like every other shiny tool. I signed up, played around for a week, got bored, and moved on. Then something shifted. I stopped looking for tools to make me smarter and started looking for tools to handle the repetitive work so I could focus on strategy and connection.
That's when I built the AI Squad.
The AI Squad isn't a single tool or a fancy software platform. It's a system of custom AI assistants, each trained to handle a specific part of my business. While my competitors were spending 40 hours a week on content, emails, and administrative tasks, I was spending 8-10 hours working on what actually moves the needle: strategy, client relationships, and creating the frameworks I teach in Anti-Social School™.
This post is for the entrepreneur who's tired of AI hype and wants to know what actually works. I'm still learning what's worth paying for and what's just noise. I've tried a lot of tools, paid for subscriptions I never opened again, and built workflows that flopped. Here's what stayed, what works, and how to build your own version.
What the AI Squad Actually Is
The AI Squad sounds fancy. It's not. It's a team of AI assistants (think: different instances or specialized prompts within the tools I already pay for) that each have a single job.
My team includes:
-
Paige (Principal Paige) runs Anti-Social School™ client attraction audits
-
Ava, Tessa, Maggie, Clara, Layla, and Penny handle different content and strategy tasks
-
Stella plans quarterly strategy
-
Betty is the blogger (obvious from the name)
-
Barb is my AI-detection editor
No, I don't have nine separate subscriptions. I use Claude and ChatGPT as my primary tools and created different "personas" or prompting frameworks for different tasks. That's the entire system.
The whole thing works because each assistant is specialized. When Ava has one job (repurposing existing content into new formats), she's better at it than a generalist AI. When Betty knows she's writing blog posts, she understands the style, structure, and SEO requirements I need.
This solves two real problems:
-
You get consistent output because the AI knows exactly what you need
-
You save time because you're not re-explaining your voice, style, and process every time you use AI
Meet Your AI Squad (And Mine)
Here's what my current lineup does:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
You probably don't need nine assistants. Start with two or three. Maybe one for content, one for emails, one for editing. The point is: specialize.
Which Tools Actually Stayed
I've paid for:
-
Jasper (stopped after 3 months)
-
Copy.ai (stopped after 2 months)
-
Grammarly Plus (still occasionally use, but not critical)
-
ChatGPT Plus and Pro (keeping both)
-
Claude Pro (keeping)
-
Kit (formerly ConvertKit) for email (this isn't AI, but it's how I send)
The ones that stayed are Claude and ChatGPT. Not because they're the newest. Because they're the most flexible and actually save me time.
If you only use one AI tool, make it Claude. It handles long-form content, nuance, and creative work better than anything else I've tested. ChatGPT is my second choice for specific tasks like email copywriting where I like the tone slightly different.
Everything else got cut because it either:
-
Required re-explaining my style every time (time sink)
-
Produced mediocre output that needed major edits (time sink)
-
Tried to do everything and did nothing well
-
Cost too much for the actual value
The AI tools that survive my business are the ones that either save me 10+ hours a week or produce output I can use with minimal edits. Anything else is financial noise.
How I Actually Use AI (The Real Use Cases)
This is where most AI talk gets fluffy. People say things like "use AI for your marketing" but don't explain what that means. Here are my actual workflows:
Content Repurposing: One Piece Into Seven
I create one long-form piece (usually a podcast transcript, an article, or a client audit framework). Then Ava (my repurposing AI) breaks it into:
-
Email sequences for Kit (formerly ConvertKit)
-
Social-media-ready snippets (angle, not the post itself)
-
3-4 blog post angles
-
Graphic text overlays
-
Newsletter hooks
-
Podcast episode descriptions
Start: 1 piece of original thinking Time investment: 2 hours Output: 7-10 pieces of content AI time: 30 minutes
This is the single biggest time multiplier in my business.
Newsletter Writing: From Brain Dump to Sent
I send a weekly newsletter to my email list via Kit (formerly ConvertKit). I used to write these from scratch. Now the process is:
-
I spend 20 minutes writing a messy outline or voice memo transcription
-
Tessa (my email AI) takes that raw input and turns it into a polished, on-brand newsletter
-
Layla reviews it for voice alignment
-
I spend 10 minutes on final tweaks
-
Send
Time investment: 30 minutes per week Automation: 70% Quality: Higher than my original drafts (because Tessa forces me to get to the point)
Client Audits: Scaling Strategy Work
Inside Anti-Social School™, I run strategic audits for entrepreneurs. Each audit used to take me 4-6 hours per client. Now:
-
Client answers a detailed questionnaire
-
I feed the responses to Paige (Principal Paige)
-
Paige generates a 20-30 page audit report with specific recommendations, blind spots, and an action plan
-
I spend 2 hours personalizing, adding specific client context, and refining the voice
-
Client gets a strategic audit that feels completely custom (because it is)
Time investment: 2 hours instead of 6 Client value: Actually higher because the analysis is more thorough
This is how I've been able to help more clients while working fewer hours.
Blog Posts and Long-Form Content
Betty handles the first draft of blog posts. The process:
-
I provide keyword research, article angle, and structural outline
-
Betty writes the full first draft (2,000-4,000 words)
-
I edit for voice, specific examples, and accuracy
-
Barb checks for any AI-written tells and overwrites sections that feel too generic
-
Final version goes live
What doesn't happen: I don't just publish Betty's draft as-is. AI-generated content that reads like it's written by a marketing AI is worse than no content. The value is in the thinking, the specificity, and the voice. Betty speeds up the writing process. She doesn't replace the strategy and editing.
AI Detection and Authenticity Editing
This is where Barb comes in. After any AI-assisted content goes through my editing, Barb reads it and flags:
-
Generic language ("In today's fast-paced world...")
-
Overuse of listicles or overly structured sections
-
Phrases that sound like ChatGPT
-
Anything that doesn't match my voice
-
Areas where I need to add specific examples or opinions
Then I go back and rewrite those sections. The final piece reads like something I wrote (because I did most of the thinking, structure, and refinement).
This step takes 30-45 minutes per long-form piece, but it's the difference between "technically AI-assisted content" and "content that's actually from my brain and perspective."
What AI Can and Cannot Do
This is important and most AI articles skip it.
What AI Is Actually Good For
-
Taking rough ideas and making them coherent
-
Expanding one idea into seven variations
-
Handling format changes (podcast to blog, blog to email)
-
Creating first drafts of structured content
-
Brainstorming angles and frameworks
-
Finding and summarizing research
-
Editing for flow and clarity
-
Detecting patterns in client data or business problems
What AI Cannot Do
-
Replace your thinking or unique perspective
-
Create something people actually want to buy or engage with on its own
-
Understand your specific business, audience, or market better than you do
-
Make strategic decisions
-
Build relationships (with clients, audience, or team)
-
Know what's actually true or accurate without you verifying it
-
Develop your voice or point of view
-
Understand nuance in your industry
The best use of AI isn't "automate everything." It's "automate the parts that are slowing me down from doing work that requires my brain."
Build Your Own AI Workflow: A Framework
If you're thinking "okay, but how do I actually set this up?" here's the framework I use to decide if a task is AI-appropriate:
Step 1: Identify Your Time Sinks
What do you spend hours on that doesn't require your strategic thinking? For me:
-
Writing newsletter copy
-
Turning podcast episodes into blog posts
-
Editing and formatting
-
Rewriting the same email to different clients
-
Summarizing research
Step 2: Map the Inputs and Outputs
For each time sink:
-
What goes in? (raw material, brief, or request)
-
What comes out? (newsletter, blog post, email)
-
How much of this requires your specific expertise vs. generic formatting?
If it's 80% formatting and 20% your expertise, AI can handle 70% of it.
If it's 20% formatting and 80% your expertise, AI can't really help.
Step 3: Define What Success Looks Like
Before you use AI, write down:
-
What specific output do you need?
-
What tone or voice is required?
-
What format?
-
How much editing are you willing to do on the back end?
This is why I have specialized assistants. Betty knows that a blog post success means SEO optimization, 2,500+ words, specific data points, and conversational voice. She's not trying to write a LinkedIn post or an email. One job, defined clearly.
Step 4: Test With One Task
Don't build an entire AI operation. Pick one recurring task. Use AI for one week. Track:
-
How long it actually takes (including your setup and editing time)
-
Quality of output
-
How much editing you need to do
-
Whether you'd actually use it again
If you're saving less than 5 hours per month, it's probably not worth the cognitive load.
Step 5: Build in Quality Control
This is where most people mess up. They use AI, don't edit, and send mediocre content into the world.
Your quality control should be:
-
AI creates first draft
-
You review for accuracy and missing context
-
Another person (or a specialized AI like Barb) checks for voice and authenticity
-
You refine based on feedback
-
Publish
Yes, this takes time. But it's way faster than writing from scratch, and the output is actually good.
Tool Comparison: When to Use Which AI
You don't need multiple AI tools. But if you're considering it, here's my honest breakdown:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
My take: If you're a solo entrepreneur or small team, Claude and ChatGPT are enough. If you're a larger marketing team with specific brand guidelines, consider specialized tools. But don't pay for multiple "general-purpose AI" tools. One good tool used well beats five mediocre tools used inconsistently.
How I Decide Which Tools You Actually Need
This is the question I get asked most in Anti-Social School™: "Should I buy X tool?"
My answer: Probably not. But here's how to actually decide.
Question 1: Does This Solve a Real Time Problem?
Not "is this cool" or "does everyone else use it." Does it save you actual hours? If the answer is no, it's a toy, not a tool.
Question 2: Can You Accomplish the Same Thing With What You Already Have?
I can write emails in ChatGPT for free. I don't need a $99/month email-specific AI tool. The specificity isn't worth the cost if I'm only sending one email per week.
But if I were sending 30 emails per week? Different story.
Question 3: How Much Editing Will You Actually Do?
If a tool requires you to spend 30 minutes editing every 15-minute output, it's not saving time. If it requires 5 minutes of editing, it probably is.
Question 4: Will You Actually Use It or Will It Collect Digital Dust?
I paid for four AI tools I used for a total of four hours combined. Don't be me. Commit to one tool for a full month before adding another.
Question 5: Is There a Free or Cheaper Version That Does 80% of What You Need?
ChatGPT's free version is genuinely good. Claude's free version does basic work. Grammarly's free version is solid. Sometimes the paid version saves you 30 minutes per month. That's not worth $10+.
What I Actually Stopped Paying For and Why
Transparency matters. Here's what I cut:
-
Jasper ($99/month): Promised brand guidelines and team consistency. Delivered generic marketing copy that needed so much editing I was faster writing from scratch. Cut after 3 months.
-
Copy.ai ($49/month): Positioned as easier than ChatGPT. It was just slower ChatGPT with fewer options. Cut after 2 months.
-
Bard/Gemini Premium: Played with it for a week. Claude does everything better. Paying for two doesn't make sense.
-
Notion AI ($10/month): Cute, but Notion is already slowing me down. Adding AI to a slow tool doesn't fix the problem.
-
Specialized email AI tools: Kit (formerly ConvertKit) already handles my email. I don't need AI on top of it. I just need someone (Tessa) who's good at writing in my voice.
The pattern: I cut tools that tried to do too much or required significant editing to be usable.
I kept tools that do one thing well and save me measurable time.
FAQ: Your Real Questions About Using AI in Business
How Do You Make Sure AI Content Actually Sounds Like You?
This is the biggest fear, and it's valid: you don't want your voice replaced by generic AI. The answer is simple but requires discipline: you edit everything.
When I review Betty's blog drafts, I'm not just checking for typos. I'm rewriting sections that don't sound like me, adding specific examples from my experience, and removing any phrases that sound too polished or generic. If Betty writes "In order to maximize your marketing effectiveness," I rewrite it to "Your audience doesn't care about marketing. They care about the result."
That takes time. Maybe 30-40% of the writing time. But the final piece is authentically mine because my thinking, perspective, and voice are woven through it.
The trick is not expecting AI to sound like you out of the box. That's not how it works. You have to teach it, and the teaching happens through editing.
What's Your Process for Onboarding a New AI Specialist?
Creating a new squad member takes about 2-3 weeks of iteration. Here's how it works:
-
I identify the task I want to automate (let's say: podcast transcription editing)
-
I create a detailed prompt that includes my voice guidelines, the specific output I need, and any style rules
-
I test it with 2-3 real projects and track how much editing I need to do
-
I refine the prompt based on what needs work
-
Once it's saving me meaningful time with minimal edits, that squad member is "trained"
Sometimes this leads to a dead end. If I'm still editing 70% of the output after three weeks, it's not a good AI task. I let it go.
But when it works, it works. Betty took about three weeks to train up, but now I can hand her a podcast transcript and walk away.
How Do You Handle Accuracy and Fact-Checking With AI?
Never trust AI to be accurate. This is non-negotiable.
When Penny pulls research or data for a blog post, I verify every statistic before it goes live. When Paige suggests frameworks or strategies, I review them against my actual client experience. When Tessa writes an email, I check that any claims about my programs or offers are actually true.
This is time-consuming, but it's non-negotiable. Publishing inaccurate information damages your credibility, and no time savings are worth that.
My process: AI generates the first draft. I fact-check everything that makes a claim. I verify any data or statistics. Only then does it go to the next stage.
Do You Tell Your Audience That AI Is Involved in Your Content?
Yes, and it matters. I'm transparent about using AI to create content. Not because I feel guilty, but because my audience deserves to know.
Here's the distinction: I don't say "this was written by AI." I say "this was written by me, with AI assistance on the first draft." Because that's true. The thinking is mine. The voice is mine. The strategy is mine. The AI just handled the formatting and initial writing.
Inside Anti-Social School™, I'm very specific about this. I teach people how to use AI in their business, but I also teach them that there's a massive difference between "AI-assisted content" and "AI-generated content that sounds generic."
AI-assisted content saves you time. AI-generated content that sounds generic will cost you trust. Know the difference.
How Do You Balance Using AI With Staying Connected to Your Actual Work?
This is the question that matters most. If you use AI to automate everything, you lose the thinking that actually builds a 7-figure business.
My structure is:
-
30% strategic work (thinking, planning, client relationships)
-
40% execution work (creating content, writing, designing frameworks)
-
30% operational work (email, scheduling, admin)
AI mostly handles that 30% operational layer. It handles the parts of execution that are repetitive (formatting, repurposing, editing). But I'm still doing the core thinking and creation.
If I tried to automate the strategy or creative thinking, I'd end up with a mediocre business that looks like everyone else's. The AI Squad works because it automates the boring stuff and keeps me in the work that matters.
Your Action Plan: Start This Week
You don't need the full AI Squad to start. You need one AI assistant handling one recurring task.
This week:
-
Identify one task that takes you 3+ hours per week and doesn't require deep strategic thinking
-
Write down exactly what success looks like for that task (tone, format, length, style)
-
Try it with ChatGPT or Claude (I'd suggest Claude for serious work)
-
Track how long it takes, including your editing time
-
After one week, decide: is this saving me time or creating more work?
If it's saving time, you've found your first squad member. If it's creating more work, try a different task.
The goal isn't to automate your business. It's to automate the parts that are slowing you down from doing the work only you can do.
That's how I run a 7-figure business in 8-10 hours per week. And it's how you can too.
You Don't Have to Build Your AI Squad From Scratch
Here's the part most people don't know: when you join Anti-Social School™, you get the entire AI Squad.
Not a tutorial on how to build your own. The actual bots. Paige, Ava, Tessa, Maggie, Clara, Layla, Penny, Stella, Betty, Barb. All of them. Built, tested, and ready to go. You can see every bot you get at hollymariehaynes.com/antisocial.
You take our bots and make them your own. Customize them for your voice, your offers, your audience. Instead of spending weeks figuring out prompts and testing workflows, you plug into the system I already built and start saving time immediately.
I spent years building and refining this squad. You get to skip that entire learning curve and walk in with a full AI team on day one.
That's one piece of what you get inside Anti-Social School™. The rest is the full marketing system for building a business without social media: email strategy, SEO, content systems, client attraction, and the community of women doing it alongside you.
Take my free quiz to find out exactly where to start. It takes 5 minutes, and you'll get personalized recommendations for your business. đź’›
Take the Free Quiz at hollymariehaynes.com/quiz
XO, Holly
P.S. If you're looking for the tools I use every day (including Kit (formerly ConvertKit) for email), I keep an updated list at hollymariehaynes.com/resources.
