Why I Licensed Wiley's Veterinary Library and SNOMED for an AI Platform
And why the difference between a veterinarian who codes and a veterinarian who hires a coder is bigger than most people think.
It's 2 a.m. and I'm standing in the treatment area of a 24-hour emergency hospital in the Houston area. I've just finished stabilizing a GDV. The dog is prepped for surgery. The owner is in the lobby, terrified, trying to understand what gastropexy means and whether her dog is going to survive the night.
And somewhere in the back of my mind, while I'm gowning up, I'm also thinking about a database schema.
That is the reality of building VetGeni. I don't build it from a co-working space in Austin. I build it between shifts. I build it because I'm still in it, four shifts a week in the ER, every week, and because the problems I'm solving are problems I lived again twelve hours ago.
This post is about two decisions I made early on that I believe define VetGeni's future, and that I think matter for any veterinarian evaluating AI tools. One is about where the clinical knowledge comes from. The other is about who is building the thing in the first place.
What a Wiley publishing license actually means
Most AI tools in veterinary medicine are built on top of large language models. GPT, Claude, Gemini, pick your flavor. Those models are trained on whatever text was available on the internet. That includes a lot of veterinary information. Some of it is good. Some of it is from a 2011 forum post where someone misremembered a drug dose.
When an AI scribe generates a SOAP note, it is pulling from that training data. When it suggests a differential or a treatment plan, same thing. And here is the part that should bother every clinician: you usually cannot see where it got the answer.
That is the problem I refused to accept.
VetGeni holds a commercial license with Wiley Publishing for their veterinary reference library. That means the clinical content in our platform, the references behind the differentials, the treatment protocols, the drug information, the toxicology data, comes from the same publisher that produces textbooks sitting on the shelf in hospitals and veterinary schools.
This is not a marketing line. This is a licensing agreement. Wiley's content is integrated into VetGeni's clinical decision support, and when the AI generates a recommendation, the source is visible. You can see the reference. You can check it. You can disagree with it. That is the whole point.
I did not build a tool that asks you to trust AI. I built a tool that shows you why the AI said what it said, then lets you make the call. That is the standard behind our AI veterinary scribe and clinical decision support platform.
What SNOMED means for a veterinary platform
If Wiley is about what the platform knows, SNOMED is about how it thinks.
SNOMED CT, the Systematized Nomenclature of Medicine Clinical Terms, is one of the most comprehensive clinical terminology systems in the world. In the veterinary world, Virginia Tech's Veterinary Terminology Services Lab has played a major role in the veterinary extension of that standard.
In plain terms, SNOMED gives each clinical concept a unique, standardized code. Acute renal failure in Conroe means the same thing as acute renal failure in a referral hospital in Sydney. Not approximately. Exactly. The code is the code.
For an AI platform, that matters enormously. When VetGeni maps a diagnosis to a SNOMED concept, it is not just labeling a note. It is connecting that case to a structured, internationally recognized clinical ontology. That means the data becomes computable, searchable, interoperable, and useful for more than today's note.
It also changes what good AI can do. Structured terminology makes it possible to build better decision support, better reporting, and better long-term clinical knowledge systems than free text can support on its own.
A concrete example of why that matters
Say a clinician enters a differential diagnosis of immune-mediated hemolytic anemia in VetGeni. That can map to a specific SNOMED concept with formal relationships to associated pathologic processes, clinical findings, body systems, and expected data points.
That structure matters because it gives the platform a framework for reasoning. It makes it easier to organize a case correctly, connect the right references, and support a more coherent treatment plan. It is not just pattern matching on likely text. It is clinical structure.
Most AI scribes in veterinary medicine do not use structured terminology in a meaningful way. They generate free text. That free text may look good today, but it does not create structured clinical data you can aggregate, analyze, or build research on later.
I wanted VetGeni on the right side of that divide from day one.
The difference between a veterinarian who codes and a veterinarian who hires a coder
This is the part of the conversation nobody in the industry seems to want to have, so I am going to have it.
There are a lot of strong veterinary AI products on the market right now. Some are built by talented software teams who spent real time with veterinarians, observed workflows, and designed products around clinical needs. I respect that work. It is not easy, and the good ones are doing it thoughtfully.
But there is a difference, a structural difference, between a product built by a developer who consulted with a veterinarian and a product built by a veterinarian who writes code.
When a non-veterinarian developer builds an AI scribe, they interview vets. They shadow in clinics. They learn the flow of a visit. The model they build is often directionally correct. But it is correct the way a map is correct. It shows the roads. It does not tell you which ones flood in March.
When I am building VetGeni at 5 a.m. after an ER shift, I am not building from a mental model. I am building from the shift I just worked. I know that the SOAP note for the blocked cat I saw at midnight has to tolerate an incomplete history because the owner was panicking and I was already unblocking the cat before we were done talking. I know that the physical exam flow needs to be fast because sometimes six systems are normal and two are the whole case. I know that discharge instructions need to land at a reading level real clients can use because I have watched too many tired, overwhelmed owners leave at 3 a.m. carrying a lot more than a handout.
A developer who has been told those things knows them as requirements. I know them as last night.
It is not just workflow. It is architecture.
The deeper difference is not in UX polish. It is in data modeling and system architecture.
When I decide how to structure a treatment protocol in the database, I am not asking a veterinarian what fields should exist. I already know that a DKA workflow needs to support fluid recalculations at intervals, electrolyte rechecks on a nonlinear cadence, insulin CRI adjustments tied to glucose curves, and the treatment of the underlying disease process, all without forcing a doctor into a rigid template that does not match how the case is actually managed.
That is not a workflow question. That is a data modeling question that requires clinical judgment to answer well. When the veterinarian and the developer are the same person, that decision gets made in one step instead of ten.
I have a computer science background. I have been writing code for decades. I went to veterinary school at Texas A&M at 40 and graduated in 2023. Before that I co-founded an oilfield trucking company and worked professional baseball, including Major League games. I have lived a few different lives. But the design advantage behind VetGeni is simple: I understand clinical medicine because I practice it, and I understand software architecture because I build it.
That is not a resume line. It is a product advantage. It shapes how the knowledge graph is organized, how SNOMED concepts are mapped, how Wiley content is surfaced at the point of care, and how a SOAP note or discharge workflow actually moves from conversation to structured output.
Why this matters for your practice
If you are a practice owner or clinician evaluating AI tools, this is the practical takeaway.
- Source transparency. When VetGeni generates a differential list or treatment protocol, the goal is not to ask you to trust a black box. The point is to anchor the output to licensed references and let the veterinarian apply judgment.
- Structured clinical data. SNOMED-linked concepts make your case data more usable than free text alone. That matters for reporting, clinical consistency, and long-term research value.
- Workflow fidelity. Because VetGeni was designed by someone still working ER shifts, the workflow decisions are not theoretical. They are driven by real use under real time pressure.
- Architectural coherence. When the clinician, product designer, and engineer are the same person, there is no game of telephone between problem and implementation.
What I am not saying
I am not saying non-veterinarian founders cannot build great veterinary software. They can and they have. The category is full of talented people doing serious work, and the industry benefits when multiple teams push each other to get better.
What I am saying is that the combination of a working clinician, a licensed veterinary reference library, a clinical terminology standard, and someone who can actually write the code is rare. That combination produces a different kind of product. Not better in every demo. Not louder in every headline. But structurally different in a way that matters over time.
What VetGeni is built on
The AI scribe category is crowded. Everybody generates notes. Everybody promises to save time. Everybody has a free trial. The question that separates a tool from infrastructure is simple: what is this thing actually built on?
VetGeni is built on Wiley-licensed veterinary references. It is built on SNOMED-backed clinical structure. It is built by a veterinarian who still works in the ER and still writes code. And it is priced to be usable, not just impressive in a pitch deck.
If that combination interests you, start a 14-day free trial. No credit card. Just your email.
And if it does not interest you yet, that is fine. Keep practicing. Keep evaluating. Keep asking vendors where their clinical content comes from. That question alone will tell you a lot about what you are actually buying.
About Chris Tiller
Chris Tiller is an emergency veterinarian in the Houston and Conroe, Texas area and the founder of VetGeni, an AI-powered clinical intelligence platform for veterinary medicine. He works ER shifts every week, graduated from Texas A&M College of Veterinary Medicine in 2023, and has a computer science background spanning decades. He is also presenting on AI in veterinary education at the AAVMC Annual Conference in April 2026 alongside Dr. Karen Cornell of Texas A&M.