Yes, AI Is Part of My Workflow. No, That Doesn't Mean You Can Skip Hiring Me.

Let me just say this upfront: yes, I use AI in my workflow. Regularly. And I'm not going to pretend otherwise just to make a case for hiring me. 😄

But there's a specific way I use it, a framework that took a while to figure out, and it's very different from what most people are doing when they try to repurpose their own content with AI tools and wonder why the results feel off.

There are actually two separate problems worth talking about here. The first is what AI can and can't do in a repurposing workflow. The second, which I think matters even more, is why so many people struggle to repurpose their own content at all... regardless of what tools they're using.

AI is fast. It's excellent at certain things: generating structural options, producing first-draft titles to react to, processing long transcripts quickly, flagging timestamps, suggesting angles I might not have considered. I use it for exactly those things.

What I never let AI do is produce the final content without edits. Here's why.

AI captures what was said, but not HOW you say it. There's a difference between the words in a transcript and the particular rhythm of how someone thinks and speaks. The framing that only comes from a specific perspective. The word choices that signal a specific point of view to someone who knows the work. AI flattens all of that into something grammatically correct and completely generic.

For a therapist, a consultant, a coach, a photographer... someone who has spent years building a credible, recognisable presence... generic content chips away at what they've built. The clients they're trying to attract are perceptive. They can tell when something sounds like the person they've been following and when it sounds like a reasonable approximation of that person. That gap matters. 🙅

AI also doesn't know what matters strategically. It doesn't know that the thing your guest said in minute 34 is the most interesting moment in the entire episode, because it contradicts what most people in your field believe. It doesn't know to lead with that. It produces output based on transcript analysis, not on an understanding of what your specific business needs to communicate to your specific audience right now. It's fast, yes. It's also very confident about things it has no business being confident about.

So the framework is: AI handles the heavy lifting on the processing end. I handle everything that requires judgment, voice, and context. The two jobs are different, and they need to stay separate.

This is the part I think it's the more interesting problem.

A lot of people try to repurpose their own podcast content. They listen back, they open a doc, plug it into ChatGPT or Claude… and either nothing comes out, or what comes out feels flat and they can't figure out why. Some spendan hour editing generic output into something usable, which defeats the whole point.

The issue, in most cases, is that they're too close to the material.

When it's your own content, you're inside it. You know exactly what you meant when you said something, which makes it really hard to hear how it actually lands to someone encountering it cold. You skip over the moment that would stop a stranger mid-scroll because to you it's obvious, you've thought about that idea a hundred times. You don't notice the framing that would resonate with your ideal client because you're not thinking about it from their perspective, you're thinking about it from yours. You miss the gold entirely because you're too familiar with the mine. 🪙

There's also the pattern problem. Every person who speaks regularly develops verbal habits they're completely unaware of. Words they overuse. Sentence structures they default to. Ways of framing things that have become invisible to them because they've said them so many times. A skilled editor catches these. The person speaking never does, because to them, those patterns feel like normal thinking.

I've sat in on enough content reviews with clients to know that the moment someone hears their own repurposed content read back to them by someone else, they immediately hear things they never would have caught themselves. "I say that all the time, don't I." Yes. You do. And it's actually interesting... it's just been diluted by repetition and needs to be used deliberately rather than accidentally. 😅

Then there's the overthinking. When you repurpose your own material, you second-guess everything. Is this too niche? Too basic? Did I already cover this? Is this the right angle or should I pull from a different part of the episode? The person who recorded the episode can't evaluate it with fresh eyes because… they don't have fresh eyes. The whole thing becomes a decision paralysis spiral, and at some point the doc just gets closed and the episode disappears like every other one.

Think of it as a proximity problem. You're too close to see clearly.

When someone who isn't inside your head listens to your episode, they hear it the way your audience hears it. They notice the moment that landed differently than you intended. They catch the phrase you use constantly that you've stopped noticing. They pull the insight you buried in the middle of a tangent, the one you were clearly excited about but then moved past, that was actually the most interesting thing in the episode.

They also know which format that insight belongs in. Whether it's a LinkedIn post or a newsletter or an Instagram caption. If it needs more context or less. Also if the framing works for the audience you're building or needs to shift slightly. That judgment comes from being outside the content.

AI can speed up a lot of the process. An outside human who understands your voice, your audience, and your business goals is the part that actually determines whether the output is any good.

That combination... fast processing on the structural side, human judgment on everything that matters... is what I mean when I say I have a framework. It's just having the right person doing the right parts of the job. 🤝