⚠️ Disclaimer: DiagnoseAI is a clinical decision support tool, not a replacement for veterinary expertise. All diagnoses and treatment decisions should be made by a licensed veterinary professional.
What You Need to Know
DiagnoseAI suggests — it does not write.
DiagnoseAI output does not automatically insert into the SOAP. Review DiagnoseAI's suggestions and manually document final diagnoses in the Assessment section and the final plan in the Plan section.
DiagnoseAI reads from the SOAP Subjective.
The quality and relevance of suggestions depend on the content entered in the Subjective field. The more detail present, the more relevant the output.
DiagnoseAI is a clinical aid, not a diagnosis.
Use it as a reference alongside your own clinical judgment. It should not be used as a primary method for determining a diagnosis or treatment plan.
Conversations may be reviewed for quality.
Human reviewers at Shepherd may process DiagnoseAI conversations to improve the tool. Do not enter sensitive personal information in the assistant window.
Enabling DiagnoseAI
To enable DiagnoseAI for your practice, navigate to Admin > Clinic Settings > AI Settings and toggle Enable DiagnoseAI on, then click Save.
If DiagnoseAI is still not visible within a SOAP after enabling, contact Shepherd Support for assistance.
Controlling Access to AI Settings
To control which users can enable or disable AI Settings, update the AI Settings permission within user roles at Admin > Users > Roles > Edit Role > Permissions > Admin Tab - check AI Settings.
How to Use DiagnoseAI
To use DiagnoseAI, open any SOAP note and click DiagnoseAI in the top right corner of the SOAP. Click Start Assistant to open the DiagnoseAI panel.
DiagnoseAI will analyze the patient information listed above and return differential diagnoses and suggested treatment considerations. You can continue the conversation in the assistant window to ask follow-up questions or request additional detail.
DiagnoseAI does not automatically write its output into any SOAP field. After reviewing DiagnoseAI's suggestions, manually enter your final clinical decisions into the appropriate SOAP sections — diagnoses in Assessment and the treatment plan in Plan.
Take A Look
How DiagnoseAI Works
DiagnoseAI is powered by a large language model (LLM) — a type of artificial intelligence trained on large datasets to understand and generate human-like language. Shepherd has worked with AI experts, developers, and veterinary professionals to build, train, and test DiagnoseAI specifically for use in veterinary clinical settings.
Knowledge Base and Sources
DiagnoseAI is designed to draw from a broad range of information sources, including peer-reviewed journals, veterinary databases, and reputable online resources. It is regularly updated to reflect current veterinary knowledge and best practices.
Because Shepherd is not directly partnered with any individual publication or information source, we are not able to identify the specific source referenced for any individual response. If a response cites or implies a specific source, use your clinical judgment to verify it independently.
Accuracy and Trust
DiagnoseAI has been tested and reviewed by Shepherd's internal team as well as external veterinary professionals to validate the quality of its responses. Shepherd continually develops specialized datasets to improve the tool's knowledge and accountability for the veterinary industry.
That said, DiagnoseAI has limitations and will not always get it right. Think of it the way you would a web search — a useful starting point and reference, not a definitive answer. Your feedback through the tool helps it improve over time.
What Info DiagnoseAI Uses
When generating suggestions, DiagnoseAI pulls the following information from the patient record:
Patient
Species
Breed
Weight
Sex
Medical Record
Initial complaint
History (from SOAP Subjective only)
Current medications and supplements
Information outside the SOAP Subjective — such as previous visit notes or lab results — is not included in DiagnoseAI's analysis.
Discharge Instructions
DiagnoseAI can generate Buoy's Law compliant discharge instructions based on the current SOAP. To request discharge instructions, ask DiagnoseAI directly within the assistant window — for example: "Please create discharge instructions for this visit" or "Generate at-home monitoring instructions for this patient."
Discharge instruction generation depends on the content present in the SOAP Subjective. If the Subjective is sparse or uses non-specific language, DiagnoseAI may not produce relevant discharge content. Using clear, descriptive language in the Subjective — or explicitly prompting DiagnoseAI to create discharge instructions — produces the most consistent results.
Discharge instructions generated by DiagnoseAI are not automatically added to the patient record. Copy the output and add it to the appropriate location in the SOAP or client communication.
Frequently Asked Questions
What is DiagnoseAI?
DiagnoseAI is a conversational AI assistant built directly into Shepherd's SOAP. Once open, you can ask it questions, request differential diagnoses based on the patient's presenting symptoms, explore treatment considerations, and generate discharge instructions — all without leaving the medical record. It reads from the patient record and SOAP Subjective, so the more detail present, the more relevant its responses. DiagnoseAI uses a large language model (LLM) — an AI trained on large datasets to understand and generate language — tuned toward veterinary clinical content.
Can DiagnoseAI replace a veterinarian in diagnosing illnesses?
No. DiagnoseAI is not a substitute for the expertise and clinical judgment of a licensed veterinarian. It is a reference tool designed to support the diagnostic process. All final diagnoses and treatment decisions must be made by a licensed veterinary professional.
Can I use DiagnoseAI in a Quick SOAP?
Yes. DiagnoseAI is available in both Comprehensive and Quick SOAP workflows.
DiagnoseAI is stuck loading or showing three dots — what do I do?
If DiagnoseAI opens but shows only a loading indicator and does not return results, check the following:
Does the SOAP have content in the Subjective field? DiagnoseAI requires Subjective information to generate suggestions.
Does the issue occur across all SOAPs, or only one specific record?
Try refreshing the page and reopening DiagnoseAI.
If the issue persists across multiple SOAPs, contact Shepherd Support and include a screenshot or screen recording, your browser and device, and the specific SOAP where the issue occurred.



