How to Build a Personal Brand Photo Series with Nano Banana

Hiring a personal brand photographer costs somewhere between “ouch” and “absolutely not.” Studio time, retouching, the awkward hour where you’re supposed to look “naturally confident” while a stranger adjusts your chin angle — it’s a whole production. Nano Banana, Promptyze’s name for Google’s Gemini 3.1 Flash Image generator, sidesteps all of that. You describe yourself, anchor your brand aesthetic, and generate a full portrait series in an afternoon.

The subject consistency feature is what makes this actually usable for branding work rather than just fun experiments. Keep the same character — same face, same build, same energy — across multiple scenes, outfits, and backdrops, and suddenly you have a coherent visual identity instead of a random pile of AI faces. This tutorial walks through the complete workflow: brand brief, character lock, scene library, and final export. No photography degree required.

What You’ll Achieve

By the end of this workflow you’ll have roughly 20 portrait-quality images showing the same consistent subject across LinkedIn headshot formats, editorial-style founder photos, and social-ready brand portraits. Every image shares your color palette, lighting mood, and overall aesthetic. The whole series looks like it came from one shoot — because in a way, it did.

What You Need Before You Start

You need access to Nano Banana, which means access to Gemini’s image generation capabilities. The fastest route is the Gemini app at gemini.google.com — image generation is available directly in the chat interface. For more control over generation parameters and batch workflows, Google AI Studio (aistudio.google.com) gives you API-level access without writing a line of code. If you’re integrating this into a production pipeline, the Gemini API via standard REST calls or Vertex AI on Google Cloud both support image generation programmatically.

You also need a clear brand brief before you touch a single prompt. Vague inputs produce vague outputs. Spend 10 minutes writing down: your subject’s physical description (specific, not generic), your brand’s primary and secondary colors, the mood you want (approachable vs. authoritative vs. creative), and three real-world visual references you admire. That brief becomes the spine of every prompt in your series.

Step 1 — Write Your Character Anchor

The character anchor is the description you’ll paste into every single prompt. It’s what keeps your subject consistent across scenes. Vague descriptions produce character drift — the person in image 12 starts looking noticeably different from the person in image 1. Specificity is your lock.

A strong character anchor covers: gender, approximate age, specific hair color and length, eye color, build, and one or two distinctive features. Avoid generic descriptors like “attractive” or “professional-looking” — those are meaningless to the model. Describe like you’re filing a police report, but politely.

Here’s a template anchor for a founder persona:

South Asian woman, early 30s, shoulder-length dark brown hair with subtle highlights, warm brown eyes, medium build, small silver stud earrings, natural makeup, confident and approachable expression

Write your own version and save it in a text file. You’ll prepend this to every prompt below. Never vary the wording — consistency in the text produces consistency in the output.

Step 2 — Define Your Brand Visual System

Before generating anything, translate your brand into visual language the model understands. Brand colors become lighting descriptions and wardrobe choices. Brand values become lighting mood and environment. “Bold and trustworthy” becomes “dramatic directional lighting, deep navy backdrop.” “Warm and approachable” becomes “soft golden-hour window light, warm cream tones.”

Write three to five “scene modules” — short descriptions of environments that match your brand. A tech founder might use: minimal white studio, modern open-plan office, rooftop city view, cozy coffee shop with exposed brick. A wellness coach might use: bright Scandinavian interior, outdoor garden setting, soft-lit yoga studio. These scenes rotate through your series while the character stays fixed.

Step 3 — Generate Your LinkedIn Headshot

Start with the workhorse image: the LinkedIn headshot. Clean background, face forward, professional but not stiff. This one gets used everywhere — email signatures, about pages, speaker bios.

[YOUR CHARACTER ANCHOR], wearing a tailored navy blazer over a white shirt, photographed from the chest up, looking directly at camera with a slight confident smile, soft studio lighting with subtle shadow, clean light gray background, shallow depth of field, professional headshot, photorealistic, 4K

The “photographed from the chest up” framing instruction is important — without it you’ll get unpredictable crops. “Shallow depth of field” keeps the background pleasantly blurred without being dramatic. Generate three to four variants of this one and pick the strongest.

Pro tip ✅

Add “looking directly at camera” to every headshot prompt. Eye contact in a profile photo consistently reads as more trustworthy than a three-quarter gaze, and the model needs explicit instruction to nail it reliably.

Step 4 — Build the Founder Editorial Series

These are the photos that go on your About page, press kit, and feature articles. More dynamic than headshots — some looking off-camera, some mid-gesture, some environmental. Generate at least six images across two or three of your scene modules.

[YOUR CHARACTER ANCHOR], wearing a structured dark green blazer, seated at a minimal wooden desk with a laptop open, natural light from a window to the left, looking thoughtfully off to the right, editorial portrait photography style, warm neutral tones, photorealistic, 4K

[YOUR CHARACTER ANCHOR], standing in a modern open-plan office, arms loosely crossed, slight smile, looking at camera, large windows with city view behind, overcast soft daylight, three-quarter shot from waist up, editorial brand photography, photorealistic, 4K

[YOUR CHARACTER ANCHOR], in a coffee shop with warm ambient lighting and exposed brick wall, holding a coffee cup in both hands, mid-conversation expression, candid editorial style, warm golden tones, shallow depth of field, photorealistic, 4K

Notice each prompt keeps the character anchor identical but varies the wardrobe color, environment, and pose cue. The wardrobe should stay within your brand palette — if your brand colors are deep green and warm gold, dress your subject accordingly across the series.

Pro tip ✅

Run each scene prompt twice and compare the outputs side by side. Nano Banana’s outputs vary meaningfully between generations — you’re not looking for the first result to be perfect, you’re looking for the best result across a small batch.

Step 5 — Social Media Portrait Variants

LinkedIn and Instagram have different visual languages. LinkedIn headshots trend formal and frontal. Instagram brand portraits trend warmer, more lifestyle-adjacent, with more context in the frame. Generate a set specifically sized and styled for social.

[YOUR CHARACTER ANCHOR], wearing a cream linen shirt, standing outdoors in a bright Scandinavian-style garden, dappled sunlight, looking at camera with relaxed smile, vertical portrait format (4:5 ratio), warm and airy aesthetic, lifestyle brand photography, photorealistic, 4K

[YOUR CHARACTER ANCHOR], at a minimal white desk with a small plant and open notebook visible, working on a laptop, overhead angle, flat lay inspired composition, soft diffused light, clean and modern brand aesthetic, photorealistic, 4K

[YOUR CHARACTER ANCHOR], close-up portrait, warm directional light from the right, soft shadow on left side of face, slight natural smile, looking directly at camera, solid warm terracotta background, tight crop from shoulders up, social media profile photo format, photorealistic, 4K

Note 💡

All images generated by Nano Banana carry Google’s SynthID watermark — an invisible, cryptographically embedded signal that identifies the image as AI-generated. It won’t affect how the image looks on screen or in print, but it means the image is identifiable as AI-generated if someone runs it through a SynthID detector. For personal brand use this is worth knowing, not worrying about.

Step 6 — Text and Overlay Variants

Nano Banana handles text rendering noticeably better than most AI image generators — which makes it useful for generating quote cards, announcement visuals, and branded content tiles directly, rather than adding text in a second tool.

[YOUR CHARACTER ANCHOR], half-body portrait on a solid deep navy background, professional and confident, looking at camera, clean text area on the right side of the frame reading "Founder & CEO", modern sans-serif typography, brand photography with graphic design elements, photorealistic, 4K

[YOUR CHARACTER ANCHOR], small figure centered against a large minimal light gray background, surrounded by generous white space, editorial magazine cover composition, confident standing pose, arms at sides, looking at camera, photorealistic, 4K

The generous negative space variant is particularly useful for LinkedIn banners and article headers where you need room to place a headline or copy over the image without covering the subject’s face.

Pro tip ✅

When you need a specific text rendering result, spell out the text character by character in the prompt: “text reading capital-F capital-O capital-U capital-N capital-D capital-E capital-R”. Tedious, but it dramatically reduces garbled letter outputs.

Step 7 — Editing Pass and Series Coherence Check

Generate your full batch — aim for 25 to 30 images across all categories — then do a ruthless editing pass. Lay all the keepers in a grid (any basic image viewer or Canva works for this) and check for character drift. If a handful of images show a noticeably different face shape or hair length, cut them rather than trying to fix them. The goal is a series where every image looks like the same person, not a series where most images do.

If you spot a scene you love but the character feels slightly off, regenerate that specific scene two or three more times rather than accepting a drift. Consistency is the entire point of this workflow — one jarring outlier undermines the whole series.

Warning ⚠️

Don’t mix Nano Banana outputs with real photographs in your brand portfolio without disclosure. The visual gap between AI-generated portraits and real photography is closing fast, but combining them in a press kit or media contact page without labeling them as AI-generated creates a transparency problem. Keep your AI series labeled, or use it as a standalone visual system.

Step 8 — The API Workflow for Bulk Generation

If you’re doing this for multiple clients or want to automate your own series refresh, Google AI Studio lets you run image generation through the Gemini API without a single line of code in the interface itself. Paste your character anchor once as a system-level context, then queue your scene prompts in sequence. You get faster iteration and can save your prompt library as a reusable template.

For production-scale work — a branding agency running this for dozens of clients, say — Vertex AI on Google Cloud gives you programmatic access with enterprise rate limits and the ability to pipe outputs directly into a storage bucket. The prompts are identical; the infrastructure scales.

Pro tip ✅

Save your character anchor and your full scene prompt library in a single document the moment you find a combination that works. Character descriptions that produce consistent results are genuinely hard to replicate from scratch — treat them like production assets.

Your Complete Prompt Cheat Sheet

Here’s a consolidated reference of the six core prompt types for a personal brand series. Swap in your character anchor, your brand colors, and your scene modules, and you have a repeatable system.

HEADSHOT: [CHARACTER ANCHOR], [brand-color blazer], chest-up, direct eye contact, confident slight smile, soft studio lighting, light gray background, shallow depth of field, professional headshot, photorealistic, 4K

EDITORIAL SEATED: [CHARACTER ANCHOR], [brand wardrobe], seated at minimal desk, natural window light from left, looking thoughtfully off-camera, editorial portrait style, warm neutral tones, photorealistic, 4K

EDITORIAL STANDING: [CHARACTER ANCHOR], [brand wardrobe], standing in modern office, arms loosely crossed, looking at camera, city view through windows, soft daylight, three-quarter waist-up shot, photorealistic, 4K

LIFESTYLE SOCIAL: [CHARACTER ANCHOR], [casual brand wardrobe], outdoors in [brand-relevant setting], dappled sunlight, relaxed smile, vertical 4:5 format, warm airy aesthetic, lifestyle brand photography, photorealistic, 4K

NEGATIVE SPACE: [CHARACTER ANCHOR], centered small figure, large minimal [brand color] background, generous white space, editorial magazine composition, confident standing pose, photorealistic, 4K

CLOSE-UP SOCIAL: [CHARACTER ANCHOR], close-up portrait, warm directional light, natural smile, direct camera gaze, solid [brand color] background, shoulders-up tight crop, social media profile format, photorealistic, 4K

Why This Actually Works for Personal Branding

The photographers who do personal brand work well charge well because they’re solving a real problem: visual consistency across many contexts. Nano Banana solves the same problem at a fraction of the cost and in an afternoon rather than a scheduled shoot day. The tradeoff is that you’re working with a described likeness rather than your actual face — which matters for some use cases and genuinely doesn’t for others. Speaker bio pages, LinkedIn banners, email headers, content marketing visuals: all of these work fine with a consistent AI-generated persona, especially if you’re building a new brand rather than documenting an existing public face.

The workflow above gives you a repeatable system. Run it once, save your character anchor and prompt library, and you can refresh your brand photo series in an hour whenever your positioning shifts. That’s the real value here — not just cheap photos, but a visual system you actually own and control.

promptyze

See Full Bio