AI in Content: A Playbook for Working Smarter Without Losing Your Human Touch

There’s more noise than ever around AI, but you don’t need another “thought leader” telling you robots will run the world by 2030. What you need is a practical breakdown of how AI can actually help you do better work, faster, and still sound like a living, breathing human who understands what a cheesesteak “wit” actually means. Here’s my full playbook, straight from the WTSFest Philly stage, on integrating AI into content workflows without selling your soul.

You’re Already Using AI (Whether You Admit It or Not)

If you’ve ever run a piece through Grammarly, used a Google Doc’s autocorrect, or copied an SEO title suggestion from Clearscope or Grammarly, congrats: you’re officially an AI user. No shame in it. In fact, it’s the new normal. The only people still pretending they don’t touch AI? Either they’ve been living under a rock or they’re about to get their world rocked the minute their boss asks for “just one more blog by end of day.”

Now, let’s drop the pretense. Whether you’re using a full-on writing assistant or just letting a plugin catch your typos, AI is everywhere in our process. The stigma is gone. The question now: Are you letting the tech help you, or are you just along for the ride?

AI is not magic. A hammer doesn’t build a house. You do. So use AI the same way you’d use a power drill. Pick it up when it makes sense, put it down when you need more finesse.

The Real Content Crunch: Why We Need Better Tools

We’ve all been there. Your content calendar just doubled, but your headcount? Not so much. Your boss wants “more, more, more.” They want blogs, landing pages, emails, thought leadership, and, oh yeah, make sure every word meets that magical “quality” threshold. Move fast, and your copy slips into mediocre territory. Slow down, and you lose momentum while competitors publish five times as much.

Welcome to the content crunch.

What’s changed in the AI era is the framing. Scaling content used to be all about speed and volume. Get more out the door, hope something sticks. Now? It’s about ruthlessly protecting your standards while getting the grunt work off your plate.

Think about all the stuff you do that’s not actually writing, like digging for stats, wrangling links, reformatting lists. Those are “line cook” jobs. Your real value? The recipe, the taste, the plating. AI should handle everything that doesn’t require your brainpower, freeing you to focus on the work only you can do best: telling the story, finding the angle, adding the human spark.

My Actual AI Stack: The Three Tools I Use (And Pay For)

Look, I’m not here to sell you a 20-tool martech stack that costs more than your monthly rent. Here’s what’s actually working for me right now, day in and day out, for me. What works for me might not work for you and vice versa, and that’s okay.

Perplexity Spaces

This isn’t your average search engine. Think of Perplexity as your always-on, never-bored research assistant. Need to know the latest stats on Philly food tourism or how many people Googled “best cheesesteak near me” last month? Perplexity surfaces current info with sources, without you spiraling into a Wikipedia hole that ends with you learning about the world’s largest ball of twine.

How I Use Perplexity Spaces

• Quick fact-checks (“When did Pat’s and Geno’s open?”)
• Deep dives on new trends (What’s the latest Google update really mean for local SEO?)
• Skimming multiple sources at once so I can spend less time collecting info and more time deciding what actually matters.

ChatGPT CustomGPTs

Forget the generic “write me a blog post” prompt. CustomGPTs let you train the tool in your voice, with your banned words, and your writing quirks. Mine has digested everything: blogs, podcasts, rants about forbidden phrases. So when I tell it to draft, it doesn’t spit out some vanilla content with zero personality.

How I Use ChatGPT CustomGPTs

• Outlining tricky topics, especially when I’m too close to the subject.
• Drafting intros, hooks, and section transitions with my usual style.
• Brainstorming alternative headlines or meta descriptions that actually sound human.

Claude Projects

This is my digital editorial assistant. Think Google Drive meets your favorite proofreader. One that remembers what you hated last time and never asks where you put the internal style guide. Claude keeps everything organized, flags awkward phrasing, and checks for those cringey “AI-isms” that instantly tell a reader a bot’s behind the keyboard.

How I Use Claude Projects

• Storing project notes, drafts, and reference docs in one place.
• Running first-pass edits to catch repetitive sentence openers or copy that sounds like it’s been run through a blender.
• Tracking feedback so I can actually see my voice evolving over time.

My Workflow: From Blank Page to Polished Draft

Here’s how it all comes together, start to finish. No theoretical fluff, just the actual process that gets a client from “I have an idea” to “let’s hit publish.”

Step 1: Research (Perplexity)

First, research with Perplexity. Punch in the topic (“2024 food tourism stats Philly”), and within seconds there’s a list of current articles, industry reports, and actual numbers with sources. Want to go deeper? The “deep research” feature pulls academic papers, case studies, or hidden gems that would’ve taken an hour to find the old-fashioned way.

Result: My research folder is packed, sourced, and fact-checked before writing the first sentence.

Step 2: Outlining & Drafting (ChatGPT)

With research in hand, toss it into the CustomGPT. Because it’s been trained on my voice, the outline comes back sounding like something I’d actually write. Something sharp, a little sarcastic, never generic. The first draft? Maybe 70% done, full of personality, and not a single “modern business landscape” in sight.

Result: Instead of starting with a blank page, there’s massaging and tightening of a draft that already feels like me.

Step 3: First-Pass Edit (Claude)

Here’s where the real polish begins. Run the draft through Claude, which checks for all banned phrases, repetitive openings, and any clunky, robotic sentences. It flags facts to review and highlights spots that don’t quite fit the style.

Result: The draft is cleaner, more consistent, and most of the heavy lifting is knocked out before ever opening a new Word doc.

Step 4: Second-Pass Edit (Human)

Now it’s my turn. Read the piece out loud. Yes, really. Catch all the awkward transitions, double-check every stat, and make sure the voice is unmistakably human. Gut-check tone and ask, “Would I actually share this?” Only then does it go live.

Result: Content that’s fact-checked, human, and actually something to be proud of putting my name on.

Why Human-AI Collaboration Wins (And Where the Robot Still Sucks)

Let’s clear up the biggest misconception. AI is a killer assistant, but it is absolutely a lousy leader.

What AI Absolutely Crushes

• Surface-level research: Finds sources, stats, and trends in seconds.
• Structure: Maps logical flow, creates outlines, flags missing sections.
• Drafting: Spits out the “boring” parts like meta descriptions or product features without burning your mental energy.
• Error checking: Flags repetitive info and redundancies before they pile up.

Where Only Humans Win

• Nuance & insight: Sussing out what’s actually new or interesting instead of regurgitating Wikipedia facts.
• Storytelling: Weaving in personal anecdotes, real analogies, and industry jokes the bots will never get.
• Audience empathy: Feeling what’s really bothering readers and responding in a way that resonates.
• Voice: Knowing the difference between “good enough” and “can’t stop reading.”

Safeguards: How Not to Embarrass Yourself With AI

AI will hallucinate and do it confidently. The AI-generated drafts claiming Ben Franklin invented the Philly cheesesteak in 1752 are real. Hilarious? Sure. Publishable? Not unless you want to be roasted on X.

My personal safeguards:

• Double-check every AI source. If something sounds off or a stat seems suspicious, dig until you find the original. Assume AI is making it up until proven otherwise.
• Read every draft aloud. If you stumble or cringe, your reader will, too.
• Watch for repetitive phrasing. AI loves to use the same construction three different ways. Cut it out.
• SEO QA: Don’t trust AI to get search intent right, to naturally work in internal links, or to deliver unique insights. That’s still your lane.

How to Squeeze More From Your AI Stack

If you want the AI to sound like you, you have to give it the raw material.

Give AI a Solid Foundation

Feed your CustomGPT everything. Past blog posts, emails, podcast transcripts, even Slack rants. The more you give, the stronger and more authentic its output.

Train Out What You Hate

Keep a running list of banned phrases, analogies, and corporate buzzwords. Every time you see an AI-ism you can’t stand, add it to the list. Update your negative prompts so your AI learns not just what you like, but what you’ll never tolerate.

Update Seed Content Frequently

Your voice changes, your perspective shifts, and you get better (hopefully) every month. Make it a habit. After every big win (killer post, viral thread, standout client project), feed that content back in. Scrap anything that no longer feels like you.

Integrate AI Into Your Routine

Start with one workflow change. Maybe you let AI do research, but you handle all the writing. Or you draft in ChatGPT and edit heavily. The point: this isn’t “set it and forget it.” Keep tweaking until you find what lets you produce more, better, faster.

Stay Flexible

AI tools update constantly. Last month’s best prompt could be outdated tomorrow. Keep experimenting, keep retraining, and treat your AI stack like a living team member. One that never stops learning.

Key Takeaways (and a Challenge for You)

• AI is your assistant, not your replacement. Use it to clear the grunt work so you can focus on what actually matters.
• Quality content always needs a human touch. The voice, the insight, the realness? That’s you.
• Start small. Pick one tool (Perplexity for research, CustomGPT for outlining, or Claude for editing) and master it.
• Iterate, experiment, break things. The only way to get better is to push the limits.
• Don’t ignore AI and don’t blindly trust it. The winners are the ones who find the balance.

My Challenge:

Start today. Don’t overthink it. The only way to lose at this is to sit back and do nothing while everyone else speeds ahead.

Ready to Actually Get Started?

While you’re debating, your competition is testing, learning, and leaving you behind. Connect with me on LinkedIn. Share your stories. The good, the bad, and the “can you believe AI wrote this?” horror shows. We’re all learning together. The best insights don’t come from theory; they come from the people who are actually doing the work.

So roll up your sleeves, elbows on the counter, and start experimenting. Just don’t let the bots order your cheesesteak.

Action Steps: Try This Today

1. Pick one AI tool (Perplexity, CustomGPT, or Claude). Create a free account if you haven’t already.
2. Upload your own writing samples (past blogs, emails, decks, anything that sounds like you). Let the tool learn your real voice.
3. Train a negative prompt list: Add banned phrases, jargon, and anything that makes your skin crawl.
4. Try using AI for just one stage of your next content project: research, outlining, or editing. Start small and see where it works best.
5. Read every AI draft out loud. Mark awkward or generic spots. Rewrite anything that doesn’t sound like you.
6. Tweak, retrain, and iterate. Keep adjusting your process every time you learn something new.

When you hit a breakthrough or break something entirely, come find me. Let’s compare notes and work together to make human-led, AI-powered content we feel good about.