Photo courtesy Instinct Science For years, veterinary teams have been trying to close the gap between the work we came here to do and the tools we were given to do it with. The first wave was better practice technology: cloud software systems built around the realities of veterinary workflows, designed to be faster, safer, and more connected, putting the medical team at the center. The second wave was AI scribing. These tools began to relieve one of the most universal burdens in practice: documentation. Less time writing notes. Less time catching up after appointments. More time being present with patients, clients, and the team. I look back on the last 10 years of practice, and these are incredible wins, but I don’t think this is the end of the story. What comes next? To see it, you have to come back to the core of a veterinarian’s work. Stripped down, our job is to diagnose, treat, and communicate. Documentation supports that work, but it is not the work itself. The third wave is going to be about something bigger: whether AI can help us not only move faster, but practice better medicine while we do it. From documentation support to clinical support The SOAP note is an important part of the visit, but by the time it is written, much of the clinical work has already happened. We have taken the history, examined the patient, interpreted diagnostics, weighed treatment options, considered risks and costs, and explained the plan. That is where technology has not yet helped enough. We have great point-of-care references like Plumb’s Veterinary Drugs, Standards of Care, journals, consensus statements, specialist recommendations, internal protocols, and years of collective experience across our hospitals. However, the harder part is getting to the right information at the moment we need it. You can know the answer exists somewhere and still struggle to find it quickly between appointments. You know a reference would help support your plan, but still have to choose between looking it up now and keeping the day moving. That is where AI starts to become more interesting: not just as a documentation tool, but as a way to consistently bring clinical decision support closer to the point of care. Practice technology is converging on practice management software, AI scribes, and clinical references, where the three work together to surface relevant clinical information at the moment it is needed during a patient exam.1 If we need to double-check a dose, review a drug interaction, look up an adverse effect, get a drug safety alert, or share a client handout while working through a case, we should not have to leave the patient’s medical record open, open another tab, and start searching from scratch. The information should be available right there, where we are already working, while the patient is still in front of us. With AI, the underlying content matters more than ever There is a phrase in technology, “garbage in, garbage out,” which suggests the quality of what you get from a system depends heavily on the quality of what goes into it. That matters even more as AI makes its way into medicine. General-purpose AI tools can produce polished, confident answers in seconds, but in clinical medicine, confidence is not the same as reliability. A tool that sounds authoritative but draws from unclear, outdated, incomplete, or nonveterinary sources can create more risk than value. That is why the content behind AI-enabled clinical reference tools matters so much. Is the content strictly peer-reviewed and evidence-based? Is it kept up to date? Does it cite published data? Is it appropriate for the species, drug, formulation, and clinical context? Is it written and weighted by veterinary experts who understand the nuance of a study design, the multi-species complexity, or the practical realities of veterinary medicine? And just as importantly, is the tool designed with appropriate safety layers, transparency, and citations so veterinarians can see where the information came from and decide whether to trust it? There’s a big difference between asking an AI model to be the source of medical truth and using AI to retrieve, summarize, and cite trusted veterinary information. One asks the model to answer. The other uses AI to speed up the research phase while keeping the veterinarian anchored to reliable references. Faster access to information only helps if the information is trustworthy. Reminder: Decision support is not decision-making AI-enabled clinical tools should support decision-making, not replace it. Clinical decision support can surface relevant information faster. It can retrieve published data, summarize key considerations, highlight potential interactions, and direct veterinarians back to the original source. However, it has limits when it comes to context. Species, age, weight, comorbidities, concurrent medications, financial limitations, specialist availability, response to prior therapy, client circumstances, lack of evidence, and more can all change what makes sense. Those details vary for every patient and situation. AI can be a very helpful assistant, taking on much of the legwork of retrieving, organizing, drafting, summarizing, and citing, but the attending veterinarian still has to synthesize the information, apply judgment, communicate the plan, and stand behind the decision. The bigger opportunity Medicine is advancing rapidly. Whether you are in a shelter environment, a busy general practice, or a large specialty hospital, we all covet the same thing: a stronger foundation of clinical confidence. That doesn’t mean cookie-cutter medicine. It means trusted information is easier to reach. Experience still matters, judgment still matters, and clinical intuition and client communication are still part of what makes a great clinician. When trusted references are available within the clinical workflow and are supported by thoughtfully designed AI tools, they can help every veterinarian make more informed clinical decisions with greater support and confidence. Human medicine is already moving in this direction, bringing AI-powered clinical decision support closer to the bedside. Our physician counterparts are using AI tools designed to quickly surface the latest evidence at the point of care.2,3 Veterinary medicine is beginning to follow, but we are at an inflection point, and these tools must be grounded in peer-reviewed content and built for how we actually practice: across species, across a spectrum of care, and within the very real nuances of practical veterinary care. That’s the bigger opportunity. AI should not just help us do more work in less time. It should create more room for the parts of practice that matter most: thinking clearly, communicating well, and making the best decisions with the most current information for the patients entrusted to our care. That’s the version of AI worth building for this amazing profession. Caleb Frankel, VMD, is the founder and CEO of Instinct Science. An internship-trained emergency veterinarian, he also serves as a board member, advisor, author, and speaker at the intersection of technology and veterinary medicine. Instinct develops the practice management software Instinct EMR, AI scribing platform ScribbleVet, clinical decision support tools Standards of Care and Plumb’s, and educational resource Clinician’s Brief. References Instinct Science. Accessed July 22, 2026. https://instinct.vet/ Open Evidence. Accessed July 22, 2026. https://www.openevidence.com/ Wolters Kluwer. UpToDate Expert AI. Accessed July 20, 2026. https://www.wolterskluwer.com/en/know/ent-ind-uptodate-expert-ai
