You want a photo that looks as though a friend took it over coffee. Instead, your AI influencer sits perfectly upright, smiles into the camera and holds the cup as if they are advertising it. The café looks convincing, but the moment does not.

When that happens, it can be difficult to know what to change. Should the skin be less smooth? Should the lighting be different? Do you need a longer prompt? Before adding more detail, give the photo a small, believable reason to exist: someone is listening to a friend, pausing between sips or turning toward the window.

Let’s build a café photo, starting with your influencer’s reference image and working through a result that still feels too staged. You do not need camera expertise. You need a clear idea of who is in the picture, what they are doing and where the person taking the photo would be.

Start with the influencer you already recognize

A reference image is a picture you give the generator to show what your AI influencer should look like. Choose a clear image with the face and hairstyle you want to keep. If you still need that starting appearance, begin with the AI influencer creation guide.

If you use several references, make sure they show the same version of your influencer. Another angle can be useful; a picture with a different face or hairstyle gives you a less consistent starting point. For this exercise, choose your references before working on the café scene, so you can concentrate on the photo rather than redesigning the influencer.

In Synthetic Actor, you can generate images using a prompt together with selected reference images. Keep the same references in place for the café photo and your next attempt. Your prompt will describe the moment around that familiar appearance.

Give the photo an ordinary moment

Imagine your influencer sitting beside a café window, talking to a friend across a small table. A coffee cup is within reach. They have just heard something amusing and are looking at the friend with a slight smile. That is enough of a scene to start with.

Notice how the action gives the expression a reason. “A beautiful influencer in a café” leaves the pose, gaze and mood largely unspecified. Describing someone listening to a friend gives you something more concrete to ask for—and something you can recognize when reviewing the picture.

Keep the surroundings connected to that moment. For our example, a cup and a folded jacket on the neighboring chair are sufficient. There is no need to add a laptop, an open book, a pastry, flowers and a phone simply to make the table interesting. Decide what this person is doing, then include the objects that belong to it.

The same principle applies to the outfit. Use something that fits the influencer you have already created. If they usually wear simple casual clothing, a familiar sweater can make more sense here than introducing a glamorous new look while you are trying to create an everyday photo.

This is also a useful way to think beyond portraits. The person remains recognizable, but the image has an activity and a setting. Later, you can build another post around choosing a book or waiting for a train without having to make every picture a new formal pose.

Describe the light and the view in plain language

You can direct this photo without naming a lens or inventing camera settings. Say where the light comes from and where the person taking the photo is sitting.

For the café, ask for daylight from the window beside the table. That gives the image one clear lighting idea. If you also ask for dramatic studio lighting and a glowing outline around the influencer, you are introducing a different visual treatment. Those choices may suit a campaign image, but they are unnecessary for the quiet conversation we are building.

Next, describe how much of the scene should be visible. A waist-up view means the picture shows the influencer from around the waist upward, with enough of the table to understand where the coffee is. “Taken from the height of a friend seated across the table” explains the viewpoint in everyday terms.

The aim is to make these instructions work together: a seated person, a nearby friend, a small table and light from the window. Google's image-generation prompting guide similarly recommends describing the subject, setting, light and composition specifically. For our example, those details matter because they describe a coherent photo you can actually judge.

Bring the scene together in a starting prompt

Here is a starting prompt for the café photo. Replace the clothing description if it does not fit your influencer, and use it alongside your chosen reference images.

Create a casual café photo of the adult AI influencer shown in the reference images. Preserve their recognizable facial features, hairstyle and overall appearance.

They are seated beside a café window, listening to a friend across a small wooden table. They look toward the friend with a slight, relaxed smile. A plain ceramic coffee cup rests on the table, with one hand resting loosely beside it. Their shoulders are relaxed.

They are wearing a simple cream sweater with ordinary folds in the fabric. A folded jacket rests on the neighboring chair. Keep the background to an everyday café interior, with a few tables visible behind them.

Use daylight from the window beside the table. Show a waist-up view from the height of the friend seated opposite, with the face clearly visible and some of the table in the picture. The mood is a quiet conversation over coffee.

The opening establishes who should appear. The middle describes a moment, including where the eyes and hand should be. The final paragraph explains the light and viewpoint. Each part has a job; you do not need to finish with a string of terms such as “masterpiece,” “flawless” or “ultra-realistic.”

Before generating, read the prompt as a description of a single picture. If you cannot imagine where the person is looking or what their hand is doing, clarify that sentence. If two instructions pull in different directions, choose the one that serves the café moment.

Look for the moment you described

When the image is ready, take in the whole picture before inspecting individual details. Does it feel like someone paused during a conversation? Or does the influencer appear to be presenting themselves and the coffee to an audience?

Then compare the person with your reference. A pleasing café scene is not enough if the face has changed. Look for the features that make this your influencer, including the face shape, eyes and hairstyle. Keep that comparison beside the question of naturalness, so a less polished result does not automatically become the preferred one.

Now follow the small action in the prompt. The cup should sit on the table, the hand should rest beside it, and the gaze should make sense for someone listening across the table. Look closely enough to notice a finger merging into the cup, an awkward wrist or a table edge passing through clothing. These are specific things you can identify, rather than a vague feeling that the whole image is wrong.

You do not need to solve every kind of image problem at once. If the scene works but the hand position does not, the guide to natural AI influencer poses can help you describe that action more clearly. If the pose works but the smile seems fixed, concentrate on the expression in your next prompt.

Revise a result that still feels too staged

Suppose the first café image keeps the right face, cup and window, but the influencer looks stiff: their shoulders face the camera squarely and the slight smile reads like a pose. Keep the same reference images and revise the description of their posture and attention.

In the starting prompt, replace the sentences about looking toward the friend and relaxed shoulders with this passage:

Their body is turned slightly toward the window while their face turns back toward the friend across the table. They are listening with their mouth softly closed, as if the friend has just started speaking. Their shoulders follow the angle of the chair rather than facing the camera squarely. The coffee stays on the table, with their hand resting beside it.

This changes something you can see. “Make it more candid” would leave the desired difference open to interpretation; the revised passage specifies a turned body, a listening expression and a cup that stays out of the spotlight.

Use the revised wording in the full scene prompt and generate another image with the same selected references. You are making a new candidate to compare with the first. Keeping the rest of the scene description stable gives you a clearer question to ask: does this posture suit a quiet conversation better, while the influencer still looks like the same person?

If your prompt contains extra phrases such as “fashion editorial,” “dramatic studio lighting” or “flawless beauty portrait,” review those too. They describe an aesthetic that may conflict with this everyday scene. Remove the conflicting direction instead of adding increasingly forceful demands for realism underneath it.

Make your next photo a moment worth showing

Try the café prompt with your influencer’s references. If the image feels posed, use the revised posture and compare the two results. Choose the one that feels more like the conversation you pictured while keeping the influencer recognizable.

Then give the next photo its own small moment: looking through a book, waiting for a train or choosing fruit at a market. Keep the reference you trust and describe what is happening around it. You will have something more interesting to create—and post—than another face waiting for the camera.

How to Make Natural AI Influencer Photos