Sparring with My Synthetic Sidekick: How ChatGPT Is Helping Me Survive the Content Tidal Wave

If you work in product marketing and haven't muttered the words "I need a clone" (often mixed with a few other choice words) at least once this quarter, month, or week, are you even doing PMM right?

We used to talk about scaling with systems. Painstakingly mapping out processes. Identifying and mitigating bottlenecks. Improving efficiency and overall driving ‘continuous improvement’. And, if we were lucky and could convince our CMO to hand over some more budget, we’d hire freelancers or interns to help us manage the workload. It worked, for a while. 

Fast-forward to today and it’s not just about ‘doing more’ anymore. It’s about doing everything, faster, with surgical precision. You’re expected to produce high-quality, persona-specific messaging for CTOs, CEOs, COOs, the occasional CFO, across three verticals, in five regions, with varying levels of digital maturity. And it all has to tie back to the master narrative. No pressure, right? Welcome to the world of modern product marketing.

Why is this happening? Spoiler alert, the villain and superhero in this story are two sides of the same ‘person’: AI. 

How did we get here? 

Let’s first address the ‘why we’re here in the first place’ part. 

Demand generation tools have become more sophisticated than ever before. The vast amount of data we’ve collected (or, let’s be real, purchased) affords us extreme levels of precision and personalization when it comes to targeting. Thank you, 6sense, for that. 

But with these shiny new toys comes a shiny new burden. The content expectation curve is no longer a curve. It’s a hockey stick. A few years ago, we could get away with a few message tracks per launch. Today? It’s not uncommon to be asked for 10+ versions of a single asset: industry-specific, persona-specific, use-case-specific, psychographic-profile-of-a-left-handed-buyer-who-likes-golf specific. And that’s not even counting revisions.

So unless you have a small army of writers (spoiler: most of us don’t, especially over here in startup land), you have to find new ways to scale your brain. That’s where AI gets interesting. No, not as a replacement, but as a thinking partner. 

Meet Isabel: My Synthetic Sparring Partner

Like many of you, I’ve dipped my toes into the generative AI pool. But recently, I decided to go full cannonball. With the help of Perplexity and ChatGPT, I created ‘Isabel’. Who is Isabel? She is a CTO in the aviation industry. Opinionated. A bit skeptical of vendors. Big on ROI, allergic to marketing fluff. She’s based on real-world insights, like actual interview notes, LinkedIn stalking (the harmless kind), and a splash of industry research. I created Isabel. And then…I started talking to her.

• I asked Isabel to review a messaging framework and a pitch deck. 

• I asked Isabel what would make her care or even purchase about our solution.

• I even asked Isabel if she’d take a meeting with one of our sales reps.

What came out of these conversations wasn’t always gold, but it was always something. Sometimes Isabel surfaced objections I hadn’t thought of. Sometimes she challenged the tone. Other times, she just confirmed I wasn’t completely off the rails. And sometimes that’s all you need. 

So, what did I learn from this experience? 

The Good: Faster Feedback, Better Focus

There are a few things Isabel (and by extension, ChatGPT) did exceptionally well:

1. Immediate, Always-On Feedback. Creative genius doesn’t always strike between 9:00 AM and 5:00 PM. And as a working mom, I often find the peace and quiet of working late to be nothing short of heavenly. Need a gut check at 10:43 PM? Isabel’s there. No scheduling, no awkward Slacks or long review cycles. Just straight-up feedback when I need it.

2. Perspective Switching. I learned I could create different persons and chat with them in parallel. One minute, I’m talking to Isabel, the skeptical CTO. Next, I’m asking the opinion of Julie, a growth-minded VP of Sales in CPG. This kind of persona shifting is hard to replicate in the real world, especially with a small team, but incredibly helpful when you’re pressure-testing your narrative across buying committees.

3. Repetitive Work = Done. Once I landed the core messaging, Isabel also helped me spin it into variants. Email subject lines. Pitch scripts. Value props. Was it always perfect? Nope. But it was a great head start that I was able to riff on. 

The Meh: Not Quite Mind-Blowing Insights 

Okay, let’s not pretend this is a silver bullet approach. Here’s what I found: 

1. Memory Loss Is Real, and Annoying. If you’ve had six other chats in between your last conversation, Isabel gets a little ‘forgetful’. You can use Custom GPTs or workspace memory (when available), but it’s not flawless. On more than one occasion I had to re-teach her who she is. It was like workplace amnesia, but with more polite responses. Tip: If you’re using a synthetic persona, ask him or her who they are first, to ensure you’re still tracking. Then dive into the feedback. Otherwise, you might get really weird responses and waste your time. 

2. Sometimes Bland, Occasionally A Bit Bizarre. Not every insight Isabel gave me was profound. Some were just… obvious. Others felt like a fortune cookie (“As a CTO, I value innovation.” Oh really? How riveting). And a few bordered on "Did I just unlock a weird corner of the internet?" Remember, generative AI still hallucinates, a lot. And from what I’ve seen, it’ll probably get worse before it gets better. If you haven’t already taken the time to do the real research and you aren’t already a semi-expert in the topics you’re chatting with Isabel about, your risk of creating ridiculous messaging is pretty high. Proceed with caution. 

3. You Still Have to Think. This isn’t a replacement for human insight. AI is great at reflecting, remixing, and accelerating. But the core strategy and original messaging still has to come from you. If your input is vague, your output will be trash. It’s like the old adage: garbage in, garbage out. 

My Overall Take On Where AI Wins: Scale Without Sacrificing Specificity

The magic isn’t in having ChatGPT write your messaging. The magic is in being able to test, iterate, and sharpen your messaging quickly, especially when you’re trying to customize it for:

• Different industries (CPG vs. manufacturing)

• Different personas (CTO vs. COO)

• Different maturity levels (AI-native vs. AI-skeptical)

Isabel became my crash-test dummy for tone, objections, and resonance. And that meant I could go into stakeholder reviews more prepared. 

Tips for Sparring with Your Synthetic Persona

If you want to try this for yourself, here’s how to get started:

1. Build a Detailed Persona Prompt. Give your AI partner a backstory. Job title. Industry. Goals. Frustrations. Personality quirks. The richer the prompt, the better the interaction. And if you like your persona, be sure to save that prompt to recreate in the future when your AI gets amnesia. 

2. Treat It Like a Real Conversation. Ask follow-up questions. Challenge the feedback. Don’t just accept the first answer. AI gets better when it gets nudged.

3. Use It to Pressure-Test, Not Perfect. Don’t outsource the core. Use AI to poke holes in your thinking, not build the whole boat. In the end, the polishing needs to come from you, building on your knowledge of the topic, solution, industry, etc. 

Final Thoughts: It’s Not Magic, But It Is Useful

Look, I’m not saying synthetic personas for sparring partners are going to win you a messaging award. But in a world where PMM teams are stretched extra thin, where the demand for personalized, resonant content keeps climbing, and where every stakeholder wants their own "version", having a tool like this at your fingertips is a game-changer. It won’t replace you. You certainly won’t be sipping piña coladas on the beach while it does your work. But it can scale you. Which means maybe you’ll have more time to have that piña colada at the end of your day. 

So go ahead. Build your synthetic sparring partner. Name them. Chat with them. Argue a little. You might just find your best messaging yet hiding in a conversation with a robot. Beep boop beep.