Running a consultancy and building a personal brand at the same time sounds manageable until you're the one writing scripts, filming, editing, and captioning every single piece of content on top of actual client work. I hit that wall, and AI-assisted content production is what got me out from under it, without giving up consistency.
The actual bottleneck wasn't ideas
I had no shortage of things to say about SaaS architecture, founder mistakes, or technical takes. The bottleneck was production time. Filming myself on camera, editing footage, writing captions for five different platforms, each with its own tone and format, that's easily the majority of the time cost in content creation, and it has almost nothing to do with the quality of the underlying idea.
Where AI avatars actually fit in
I use an AI avatar for video delivery on shorter-form content, scripted by me, delivered by an avatar built to match a consistent visual identity rather than requiring me to be on camera for every single piece. This isn't about hiding who I am. It's about decoupling content volume from my personal filming availability.
The scripts are still mine, written in my voice, based on real lessons from actual client work. The avatar changes who delivers the line, not what the line says or where the idea came from.
What this workflow actually looks like
- Write the core idea and script myself. This is non-negotiable. The credibility of the content depends on it actually reflecting real experience, not generic advice generated without context.
- Calibrate the script to natural speaking pace, so the delivery doesn't sound rushed or artificially slow. This took real trial and error to get right, most default AI voiceover pacing assumes a much slower delivery than how people actually talk.
- Generate the avatar delivery in short, structured clips, rather than one long take, which keeps editing modular and makes revisions faster when something needs to change.
- Layer in platform-specific captions and formatting for each destination, Instagram, LinkedIn, YouTube, rather than posting the identical caption everywhere and hoping it performs the same way across platforms with completely different algorithms.
What I learned about the limits
AI avatar tools generally won't create a video of a real, identifiable person from a reference photo, mine included. That's an intentional safety limitation, not a bug to route around. The workaround isn't to force a likeness match. It's to use a generic description for avatar appearance and be transparent, where platforms require it, that AI is part of the production process.
For content where actual likeness matching matters, there are dedicated tools built specifically for that, at a different cost and complexity tradeoff than what I use for daily short-form content.
The disclosure question, and why I don't dodge it
Some platforms have explicit requirements around disclosing AI-generated content, and I follow them without trying to find loopholes. Beyond the compliance angle, I think it's the right call anyway. The value of the content is in the ideas and experience behind it, not in pretending a human filmed every second of it. Being upfront about the production method hasn't hurt credibility. If anything, it's a reasonable thing to be transparent about.
What actually improved once I made this shift
Consistency, mostly. I went from sporadic posting driven by whenever I had a free afternoon to film, to a repeatable production process I can run alongside actual client work without either one suffering. That consistency is what's actually built the audience over time, not any single viral piece of content.
The bottom line
AI-assisted production isn't a shortcut around having something real to say. It's a way to remove the production bottleneck between having a real idea and actually publishing it consistently. The ideas, the experience, the voice, that all still has to be genuinely yours. AI just handles the parts that used to eat all the time before the idea ever reached an audience.
Suggested internal links: Link to Day 9's AI-augmented development article and Day 19's AI roadmap piece.
CTA: Curious about the exact workflow? Follow along, I break down the process regularly.