AI Simulations vs Clinical Before-and-After Photos

An AI simulation and a clinical before-and-after gallery can both support an aesthetic consultation, but they do different jobs. One illustrates a visual direction on the current person's photo. The other documents what happened to a different patient after a real treatment.

They answer different questions

A person looking at an AI preview is usually asking, “Is this visual direction worth discussing?” A person looking at a clinical gallery is more often asking, “What kind of work has this practice documented in real cases?” Mixing those questions creates confusion.

What a simulation can do well

A bounded simulation can make language such as “subtle,” “more defined,” or “less tired-looking” concrete enough to discuss. Because it begins with the person's own portrait, it may reveal that the provider and patient were imagining different magnitudes or target areas.

That makes the image a communication artifact. It can help someone point to what they like, dislike, or want to avoid. It can also help a provider say that the requested direction is not achievable, not appropriate, or would require a different conversation. The value is not that the software has decided what will happen.

What it cannot establish

A single front-facing portrait does not establish three-dimensional anatomy, facial movement, medical history, contraindications, product behavior, technique, swelling, recovery, or healing. It cannot determine candidacy or choose a treatment. A polished output does not change those limits.

What a clinical gallery can do well

A carefully documented gallery can show the practice's real work across multiple cases. It can demonstrate the range of concerns treated, how results were photographed, and how appearance changed at a stated follow-up point. It may also help a provider discuss variation rather than presenting one idealized case.

The gallery is only as trustworthy as its documentation. Compare like views: similar angle, expression, camera distance, lens perspective, light, and framing. State the treatment and time between photographs when it is appropriate and permitted to do so. Do not let makeup, hair, a smile, or dramatically better light carry the “after.”

What it cannot establish

A real result is still one person's result. Selection bias matters: a gallery usually shows cases the practice chose to publish. The viewer may have different anatomy, goals, history, treatment plan, recovery, or response. A gallery supports a discussion about documented work; it does not forecast the next patient.

How to present both without blurring the line

1. Label before interpretation. Put “AI simulation” or “actual patient result” beside the image, not only in a distant footer.
2. Use separate visual treatments. Do not place generated previews inside the same unlabeled carousel as clinical cases.
3. Explain the job. Introduce a simulation as a preference-setting aid and a gallery as documentation of other patients' cases.
4. Keep provenance. Record permission and source for published clinical photographs. Keep generated examples synthetic or covered by specific publication rights.
5. Preserve the provider's correction role. The provider should be able to reject the simulated direction and explain why a real plan may differ.
6. Never imply an outcome match. Avoid captions that suggest the current person can achieve the adjacent clinical result or simulated appearance.

Apply a different quality check to each format

For an AI simulation, score identity preservation, treatment-area fidelity, changes outside the intended region, artifacts, repeatability, failure handling, and disclosure. The AI simulation evaluation guide provides a reusable rubric.

For clinical photos, audit consent for publication, case provenance, matching views, capture conditions, timing, edits, captions, and whether the presentation overstates what one case means. Both formats need honest context, but only the clinical pair documents a real procedure.

A practical consultation sequence

Start by asking the person what they want to discuss. If they brought or created a simulation, ask which part communicates the preference and which part does not. Conduct the provider assessment independently. Then use relevant clinical cases, if available and appropriately consented, to discuss real-world variation and the provider's experience. Document the actual recommendation separately from the generated image.

Eva's free public preview and treatment-specific routes such as the Botox simulator label outputs as illustrative simulations. Practices evaluating the broader workflow can review Eva AI Before & After and the provider-led consultation workflow.