Ambient documentation tools draft the record during the exam, so charting gets done while the visit is still fresh rather than after closing. GettyImages/PixelsEffect Ask 10 practice owners whether AI has paid for itself, and you will get 10 different answers, ranging from, “It’s transformed our documentation” to “We signed up, and I couldn’t tell you if anyone uses it.” Both responses are common. The difference between them has less to do with the software than with how the practice measured, rolled out, and adopted the tool. AI has moved quickly from novelty to fixture in veterinary medicine. A 2024 survey of 3,968 veterinary professionals, conducted by Digitail in collaboration with the American Animal Hospital Association (AAHA), found 39.2 percent of respondents were already using AI tools or software in their practice, and among those who had tried it, nearly 70 percent reported using it daily or weekly. Another 38.7 percent said they were interested in adopting AI soon.1 In other words, this is no longer a fringe experiment; it’s becoming standard equipment. However, adoption and return are not the same thing. Understanding the real ROI means getting specific about where the value shows up, how to measure it, and whether the tool is being used at all. Where AI is returning value today Clinical documentation Ambient scribe and transcription tools listen during the exam and draft the record, so the veterinarian is not typing while a client is talking or catching up on notes after closing. The return here is concrete: time. When charting moves out of the after-hours pile and into the appointment itself, doctors reclaim evenings, and records get completed while the visit is still fresh. The downstream benefit is cleaner records, which support better handoffs between doctors, technicians, and front-desk staff. Practices evaluating these tools should track charting completion time and the number of records completed the same day. Both are easy to baseline and re-measure. Diagnostic and decision support A second category uses a patient’s history and presenting signs to surface relevant considerations for the clinician, not to make the call, but to make sure nothing obvious gets missed during a busy day. The framing matters here. These tools earn trust when they behave like a well-prepared colleague who offers context, and they lose it quickly when they overpromise. For complex or chronic cases, accuracy matters far more than speed, and any responsible implementation keeps the veterinarian firmly in control of the decision. Used this way, the value is in reduced cognitive load and fewer things slipping through the cracks. That’s harder to put a dollar figure on, but real. Client communication The third area is the least glamorous and often the highest-ROI: automating and improving the flow of information to pet owners. Reminders, follow-ups, appointment confirmations, and discharge summaries consume an enormous amount of front-desk time when handled manually. Tools that draft client-ready summaries or reduce phone tag free staff for higher-value work and done well, improve the client experience at the same time. The metrics to watch: no-show rate, reminder compliance, and time spent on outbound calls. “At my office, clinical documentation gave us our first real win. Before AI, I spent too much time after clinic reconstructing detailed mobility and therapy notes for patients,” says Tom Walsh, DVM, owner, PrimePet Rehabilitation Therapy in Myrtle Beach, S.C. “Once the tool listened during visits, notes wrote themselves while we talked with owners, creating cleaner handoffs and fewer missed details from busy physical therapy days. In the AI world, documentation often is the entry point because everyone feels the drag of falling behind on records, and fixing it improves both daily life and record quality immediately.” To know the ROI, you have to know the ‘before’ Here is the uncomfortable truth about measuring AI’s return: most practices never recorded where they started. ROI is a comparison. It requires a baseline. If you don’t know your average charting completion time, your no-show percentage, your reminder compliance rate, or how many hours staff spend on the phone before you adopt a tool, you have nothing to compare against afterward. What tends to happen instead is the practice adopts something, notices that things feel a little smoother, and either credits the software with everything or, when the honeymoon fades, quietly decides it did not do much. Neither conclusion is grounded in data. The fix is unglamorous and worth the effort. Before implementing any AI tool, spend two to four weeks capturing a handful of baseline numbers tied to the outcome you care about: For documentation tools: Average time to complete a record, percentage of records finished same day, after-hours charting hours per doctor. For communication tools: No-show rate, reminder response rate, outbound call volume or minutes. For decision support: Harder to quantify, but track things like missed-charge frequency or record completeness, which often move alongside better documentation. The adoption problem The single most common reason AI investments underperform has nothing to do with the technology. It is the tool that was handed down from the top without the team being on board. Leadership evaluates a tool, sees the promise, signs the contract, and announces the rollout. The staff—who were not consulted, nor trained beyond a rushed lunch-and-learn, nor asked what problem they hoped to solve—treat it as one more thing imposed on an already full day. Within a few weeks, people revert to the old way. The subscription renews anyway. This is not resistance for its own sake. Clinical teams are protective of their time and their judgment, and skeptical of anything that promises to “transform” work they already know how to do. Documentation tools affect the medical record, which is professionally and legally the doctor’s responsibility. If a veterinarian does not trust the output or feels the tool is being done to them rather than built around them, they will not use it. The practices that get real ROI do a few things differently. They involve the people who will use the tool in the decision. They pilot with one or two willing team members—often the doctor who is pickiest about workflow, on the theory that if the skeptic likes it, everyone will. They frame AI as optional and doctor-controlled, an accelerator rather than a replacement for clinical judgment. Plus, they connect the tool to a problem the team agreed was worth solving, rather than a capability someone wanted to buy. “Team buy-in matters more than any feature set when it comes to real ROI. In my practice, I brought our front-desk lead and a senior rehab tech into the process before we rolled out client communication tools,” says Dr. Walsh. “They helped customize discharge summaries and reminders around home exercise plans and mobility progress for our patients. Because the team shaped it from the beginning, adoption was high, and we saw better compliance with less time spent on follow-up calls. It really was a painless experience.” The metric no one tracks Adoption rate is the most revealing metric in the AI conversation, and the one that practices most consistently fail to track. It is simple to define: Of the people or encounters that could use the tool, what percentage actually do? If five doctors have access to an ambient scribe and two use it regularly, your adoption rate is 40 percent, and your real per-user cost is more than double what the invoice suggests. Most software platforms surface this data if you look for it. Usage dashboards, feature-level reports, and login activity tell you whether a tool has taken root. It is the fastest way to diagnose a problem while there is still time to fix it. A low adoption rate three months in is not a reason to abandon the tool; it is a signal to revisit training, workflow fit, or team buy-in. It is worth remembering that in the AAHA survey, the professionals who adopted AI overwhelmingly stuck with it—nearly 7 in 10 used it every day or week.1 Adoption, once it takes hold, tends to last. The trick is getting there. Tracked honestly, the adoption rate reframes the ROI question. A tool used by everyone that saves each doctor thirty minutes a day is a bargain. The same tool, used by no one, is a line item you should cancel. So, the practices seeing genuine return on AI are not necessarily the ones that bought the most capable tools. They are the ones who wrote down where they started, brought their teams along, and kept an honest eye on whether anyone was using what they paid for. The technology will keep getting better. Whether it pays off in your practice still comes down to those three habits and the willingness to check the numbers, even when the answer is that the tool is sitting idle. That’s the audit worth running this quarter: not whether AI works, but whether it’s working for you. Tom Walsh, DVM, is the owner of PrimePet Rehabilitation Therapy. He co-wrote this article with the team at Shepherd Veterinary Solutions.
