The skin photos already arriving in your consultation flow: why they’re unstructured

Author: Alexander Börve

October 5, 2026

Patients already send photos of their skin concerns as part of everyday consultation flows. For instance, a rash explained in a message often comes with a phone photo attached. A pharmacist handling an online query may receive the image before they have even finished reading the description. The behaviour is already part of how people seek advice.

What often gets overlooked is what happens once that photo arrives. In many consultation flows, it remains just an attachment that waits to be opened and reviewed. Nothing structured is added to the image before the human part of the process begins.

This is not a case for changing patient behaviour. Patients are already sending the photos. The question is what the consultation flow can do with those images once they arrive.

Where these photos already show up

Skin photos can enter an online consultation through several familiar touchpoints:

Asynchronous messaging with a GP, pharmacist or nurse, where a photo is attached alongside a written description.
E-consult intake forms, where patients can upload a photo alongside their symptoms.
Symptom-reporting flows in primary care or online pharmacy settings, where an image accompanies a structured questionnaire.
Follow-up conversations, where a patient sends another photo to show how a skin concern has changed.

These are already ordinary parts of online consultation and primary care flows. Patients use photos because some skin concerns are difficult to describe clearly in words alone.

What happens to them once they arrive

Once a photo is attached to a message, form or consultation, it typically waits for the next part of the existing process. That may be a pharmacist answering a query, a GP reviewing an e-consult or another human step within the consultation flow.

The image itself does not necessarily provide anything to the patient while they wait. The patient who sent it gets no immediate sense of what the photo might show; they have to wait until that human step is completed, whether that is a few minutes later or a day or two later.

This creates a gap between submitting a photo and receiving a response. Patients may have sent the image because they want to provide more information about their skin concern, but the image remains separate from any immediate informational feedback. For platforms handling hundreds or thousands of skin photos each month, this leaves a recurring gap between what the patient wants in the moment and what the consultation flow currently provides. The photo is present, but its role in the patient experience can end at submission.

What structuring this flow would look like

Structuring this flow means adding a self-service layer at the point where the patient submits the photo. As soon as a patient submits a skin photo, they receive a ranked set of informational suggestions about what the image most closely resembles, before the submission continues through the human-reviewed part of the consultation.

That gives the photo an immediate purpose for the patient. Instead of becoming an attachment that waits in a queue, it provides something useful in the moment. It helps them understand what they may be looking at. It can also make the wait for the existing consultation process feel less uncertain.

Where Autoderm fits

Autoderm is built to power an informational skin check within an existing consultation flow. Rather than sending patients to a separate destination, a service can add the feature to a journey they are already using. The patient submits their own skin photo. They receive Autoderm’s ranked, informational condition suggestions. The self-service experience stays within the existing consultation flow, while the established human process continues alongside.

Published evidence reports top-5 suggestion accuracy of 72% across all 100 patients in a peer-reviewed primary care study (Escalé-Besa et al., 2023). A white paper, the Coachella Study (2025), reports 93%. Figures differ by study design, population and model version, so each should be read against the clinical evidence page.

Under Autoderm’s data processing agreement, Autoderm processes submitted images as a processor on the service’s behalf and strips image metadata. Images are retained for model training only where the service sends an explicit consent signal.

Autoderm also holds CE marking as a legacy Class I device under the MDD, with MDR Class IIa in progress. For a service considering this type of self-service feature, the next step is a scoping conversation rather than a plug-and-play setup. Before that, see what to verify before integrating a dermatology API. You can also create a free account and use the test console to explore image handling and response format.

Alexander Börve

Alexander Börve, MD is the founder and CEO of Autoderm. He has worked in digital dermatology since 2006 and has published seven peer-reviewed papers on mobile teledermatology and AI-assisted dermatology.