How Terav Works

Work through a clinical case with Terav

Terav is an AI clinical reasoning assistant for veterinary professionals. It brings patient information, laboratory findings, clinical discussion, and relevant literature into a case-based workspace.

Start with the information you have, add findings as they become available, and ask Terav to help examine the case. You can explore differential diagnoses, identify missing information, question an interpretation, and review the sources behind retrieved evidence. You decide what matters for the patient.

1. Create a case and add the clinical context

Create a case with the available patient details, such as species, breed, age, sex, and weight. Describe the reason for the visit and the clinical question you want to work through.

In the case conversation, you can add relevant history, examination findings, previous investigations, medications, treatments, and your working hypothesis. Plain clinical language is enough; you do not need to construct a specialized AI prompt.

For example:

Here are the history and examination findings. Help me organize the differential diagnoses, identify missing information, and explain which findings support or weaken each hypothesis.

Terav uses the patient details and available conversation context to inform its response. Make any information essential to the current decision explicit, including dates and changes since the previous assessment.

2. Add laboratory results and check the extracted data

Upload a laboratory report as a PDF or photograph. Terav extracts results for review and interpretation within the case.

Before structured laboratory interpretation proceeds, the system checks for issues such as uncertain recognition, unfamiliar test names, unit mismatches, and implausible values. If entries are flagged, you can correct them or confirm that they match the original report. Those flags must be resolved before the review is confirmed and interpretation continues.

Results that do not trigger a flag can proceed automatically. Check the extracted information against the original report, especially where a value could change your clinical decision.

For numeric results, a rule engine runs before the AI interpretation. It normalizes test names, applies supported unit conversions, compares values with available reference intervals, and looks for predefined patterns and critical values. Where a usable reference interval is extracted from the report, it takes priority over a catalogue reference.

The AI receives these results alongside the available patient context to help explain their clinical significance. The analysis depends on the information supplied and the coverage of the underlying rules and references.

3. Review the reasoning and differential diagnoses

Terav’s clinical reasoning format separates case facts from hypotheses and asks for the findings that support or weaken each explanation.

A case analysis can include:

• A summary of the information available.
• A ranked differential list with arguments for and against each hypothesis.
• Missing information that would help clarify the case.
• Diagnostic steps linked to the hypotheses they would help assess.
• Relevant limitations and uncertainty.

Hypotheses use qualitative support labels:

• Supported: several independent findings align.
• Possible: the explanation fits, but information is limited or a competing hypothesis remains.
• Unlikely: the explanation fits poorly but has not been excluded.

These labels help you examine the reasoning. They are not diagnostic probabilities, and the leading hypothesis remains open to revision. Terav is designed to support a differential diagnosis rather than issue a final diagnostic verdict.

4. Challenge the assessment

You can question the reasoning directly in the conversation. Terav also provides Challenge prompts for examining the leading hypothesis and the assumptions behind it:

• “Argue against the leading diagnosis.”
• “What is this assessment hanging on?”
• “What would change this assessment?”

These prompts ask Terav to look for contradictory findings, an observation it may have overemphasized, an alternative that deserves more attention, or information that would change its assessment.

A clinical check also invites you to compare the response with your own assessment. Add what you observed, explain where you disagree, or introduce new findings to continue the discussion.

5. Read the original evidence

Terav’s article knowledge base contains original publication text, preserved without rewriting or summarization and divided into searchable passages called chunks. Each passage retains its connection to the source article.

Through retrieval-augmented generation, or RAG, relevant passages can be supplied to the AI alongside the case context. Source links let you open the original publications and examine the evidence for yourself.

The article text is the source material; the explanation applying it to your case is generated by AI. When a decision depends on a claim, review the article’s methods, study population, limitations, and relevance to the patient.

Available evidence and retrieval coverage vary by question and workflow. A source link helps you inspect the evidence; it does not mean every statement in the answer has been independently verified.

6. Continue the case as new information arrives

Return to the same case to add findings, report investigation results, clarify the history, or discuss a change in the patient’s condition. The case keeps the conversation and attachments together so you can return to the earlier discussion.

Terav uses available case context and conversation summaries to support follow-up questions. If you want to compare particular results over time, identify the relevant values and dates in your question. Do not assume every detail from a long conversation is included in each new response.

The case workspace also supports a structured clinical note with a problem list, assessment, and next steps. Review it against the original information and your own conclusions before using it in clinical documentation.

Your clinical judgment guides the next step

Terav can help organize a case, interpret findings, explore hypotheses, and discuss diagnostic approaches or treatment options. The veterinarian assesses the patient, verifies important information, and decides which actions are appropriate.

Terav is intended for veterinary professionals. It does not replace examination, specialist consultation, or emergency assessment, and it is not a self-diagnosis or treatment service for pet owners.

For a detailed explanation of source retrieval, laboratory checks, clinician review, and their limitations, read Evidence & Safety.

Start with a case. Add the findings. Examine the reasoning.