How to Make a Business Portrait in ChatGPT From an Ordinary Selfie

A photo for a job site or a resume is usually solved the same way: studio, photographer, retouching, a week of waiting. And then it turns out you need one more shot for a presentation, in a different format and with a different background.

ChatGPT can build such a portrait from ordinary phone shots. Here is how it works in practice and where the prompt wording decides whether it looks like you.

What happens technically

ChatGPT with image generation can use uploaded photos as a face reference. You upload 3 to 5 of your shots, preferably in different lighting and angles, and ask it to build a portrait based on them. The model follows the facial features from the reference: the shape of the eyes, the nose, the lips, the oval of the face. Then, following the prompt text, it changes only what you asked for: clothing, background, light, pose.

After that, you can ask for different shots of the same face with separate prompts: an avatar for a job site, a strict portrait for a resume, a shot in an office, a shot on stage with a microphone. Each shot is essentially a separate request on top of the reference you already uploaded.

The prompt decides more than it seems

The most common mistake is asking for just a "business portrait" with no details. In that case the model picks the genre itself, and the result can drift toward glossy stylization instead of a recognizable, similar photo.

A working prompt describes both the genre and what must be preserved:

"Business portrait for a resume: strict suit, neutral light gray background, studio photography lighting, keep realistic skin texture with no smoothing and no retouching."

Two clarifications in this prompt make all the difference. "No smoothing and no retouching," so the face doesn't slide into plastic gloss that gives away an AI at once. "Neutral background," so the shot works both for a job site and for a presentation without extra details behind you.

Formats that cover different tasks

One round with the uploaded photos can be expanded into several shots for different purposes:

• - a strict portrait in a city business district for LinkedIn and job sites;
• - a shot in an office at a desk for a team presentation;
• - a shot on stage with a microphone for a speaker profile;
• - a shot with empty space along the edge for a slide cover or banner, where text goes later.

The last item saves a separate trip to a graphic editor: a composition with room for text is set right in the prompt, not cut out by hand afterward.

Where this method doesn't fit

If the goal is a photo of yourself younger, slimmer, or with different facial features, that is no longer a business portrait but stylization, and the result will naturally look less like the real face. The more you ask to change in your appearance, the further the result drifts from a recognizable shot.

For official documents (passport, visa, pass) such a portrait doesn't work for formal reasons: they need a real photo with specific technical requirements, not a generated one.

Bottom line

A business portrait from an ordinary selfie in ChatGPT solves a practical task: a photo for a job site, a resume, or a presentation without a photographer and a week of waiting. Two things decide it: a good reference (3 to 5 different photos) and a prompt that sets the genre and at the same time asks to keep the face realistic without smoothing.

I write about practical setups like this with AI in the channel «Растём с ИИ» ("Growing with AI"). A breakdown of a related topic with other generation models: how to choose an AI tool for the task.

If you need to build such a prompt system for your personal brand, you can start with a free consultation.