By skyforbes My issue wasn’t writing. I could crank out LinkedIn posts all day.
My issue was visuals. I’d finish a post and then get stuck at the image step. No recent photos, nothing that matched the vibe, and I didn’t want to use the same headshot every time.
So I wrote this prompt to turn ChatGPT into a sort of photo art director:
“You are a professional photographer and visual storyteller. I’m writing a post about: [PASTE POST]. Suggest 5 photo concepts that would visually support this message. For each concept, include:
– Setting/background (be specific)
– Outfit/style
– Expression/body language
– Lighting mood
– Overall vibe in one sentence.”
The output from that alone was gold. Instead of “generic office shot,” I’d get things like:
“You at a standing desk with a laptop, soft morning light from a side window, sleeves rolled up, expression focused but relaxed—conveys ‘in the work but not overwhelmed.’
Then I’d take one of those descriptions and drop it into Looktara, which is trained on my face and can generate photos of me specifically.
Result: A realistic, on‑vibe image for my post in about 10 seconds, without a camera.
The real win isn’t just the AI headshot. It’s using a good prompt to get specific visual directions so the image model has something meaningful to work with.
If you’re already deep into prompt engineering for text, it’s worth asking: “Can I use the same skills to direct the visuals around my content too?”