All Episodes

From Apple to Animals: The Pivot That Built a 6 Continent AI Company in 30 Months

with Yannick Bloem · CoVet

August 24, 202600:54:11

From Apple to Animals: The Pivot That Built a 6 Continent AI Company in 30 Months

0:000:00

Show Notes

Everybody says the problem with veterinary medicine is burnout. Yannick Bloem thinks the problem is paperwork.

A veterinarian finishes a full day of consults. The animals have been seen, the owners have gone home, and then the second shift starts: SOAP notes, discharge summaries, referral letters, records that have to exist by law whether or not anyone has the energy left to write them. Two hours, most nights. That is the part of the job nobody signed up for, and it is the part that quietly pushes people out of a profession they love.

Yannick Bloem spent five years as a senior software engineer at Apple building productivity software for hundreds of millions of people, and before that built a consumer startup in Switzerland that reached over 100,000 users. When GPT landed in late 2022 and he was living in San Francisco watching doors open in real time, the obvious move for someone with that resume was to chase the frontier. He did the opposite. He called a friend from Toronto whose mother is a veterinarian, and they went hunting in the least fashionable corner of healthcare they could find.

That friend is Mike Parent, now CoVet's COO, who had grown up filling out paperwork alongside his mom during school holidays. They got a rough, ugly, half-working product into the hands of about 20 beta vets through Mike's network. The vets loved it anyway. That gap between how bad the product was and how excited the users were is the entire founding signal.

Two and a half years later, CoVet is used daily across North America, Europe, and Australia, in 20 languages, by groups including Mars Veterinary Health, IVC Evidensia, CVS Group, VetPartners, Petfolk, and Small Door Veterinary. User volume grew 550% in 2025 alone. The company employs over 30 DVMs and RVTs in house, and Yannick says the team is approaching 130 people. It has raised remarkably little to get there.

The episode goes deep on three things founders keep getting wrong: that the newest frontier model is automatically the right model, that regulatory burden is a tax rather than a moat, and that you need a mega round to scale. There is also a section on the domain, which Yannick correctly describes as drip, and a section on why horses need so much more paperwork than cats.

Named Frameworks

The Two Hour Dividend

CoVet's core claim is not a feature, it is a unit of returned time. What makes it durable is that the company does not prescribe how the time gets spent.

  • The product automates the legally required record, not a nice to have.
  • Vets recover roughly two hours per day, compounding to weeks per year.
  • Some vets spend it seeing more patients, some spend it going home for dinner.
  • Because the user chooses the payoff, the value proposition survives every persona.
  • The testimonial that closes deals is about family dinner, not about throughput.

Right Model, Right Task

Yannick's answer to model selection is an internal routing layer, not a vendor bet.

  • Every task carries a speed, cost, and quality tradeoff, and they rarely optimize together.
  • Frontier models are heavily tuned for coding, which is not the veterinary use case.
  • Engineers cannot judge whether a SOAP note is clinically good, so vets run the evaluations.
  • Open weight models running on owned hardware solve the data safety question at the same time.
  • Assume the newest release is wrong for your task until your own evals say otherwise.

Highest Standard Default

Rather than mapping the minimum compliance requirement per market, CoVet built to the ceiling once.

  • Veterinary data faces materially lighter regulation than human healthcare.
  • CoVet adopted SOC 2 Type 2, ISO 27001, GDPR, HIPAA, and PIPEDA anyway.
  • Contractual guarantees with model providers block training on customer data.
  • Enterprise veterinary groups and European buyers require this bar to even take a meeting.
  • One standard means the company can enter any market without a compliance rebuild.

Seed Strapping

The capital philosophy Ryan names and Yannick confirms.

  • Strong early cash flow funds growth instead of dilution.
  • Constraint forces better decisions than abundance does.
  • First outside money came from veterinarians who had sold clinics, buying validation alongside capital.
  • Every subsequent raise happened on top of growth that had already occurred.
  • Raising from strength beats raising to reach somewhere you cannot see yet.

Two Motion Go To Market

CoVet runs bottoms up and top down in parallel rather than sequencing them.

  • Independent vet motion: classic sales reps, self serve trial, low friction.
  • Enterprise motion: consolidating groups that want one standard across many clinics.
  • The industry's consolidation trend makes the enterprise motion structurally advantaged.
  • Product quality is the wedge in both, because enterprise buyers are shopping for the best tool their vets will actually adopt.

Habit Before Hire

The student program is a distribution strategy dressed as generosity.

  • Free unlimited access for veterinary students and interns throughout training.
  • Access continues for 8 months after graduation.
  • Partnerships with schools including Colorado State University's College of Veterinary Medicine.
  • Graduates arrive at their first clinic with the product already configured to them.
  • Association member benefits, including the Montana Veterinary Medical Association, extend the same motion to working professionals.

Ugly But Useful

The founding validation test.

  • Ship something that does the basics badly to about 20 real users.
  • Ignore what they say about the interface.
  • Measure enthusiasm relative to how bad the product currently is.
  • A large gap between quality and excitement means the underlying pain is enormous.
  • That gap, not polish, is the go signal.

Founder Experiment

Run a boring invoice audit over one week. Open your last three months of expenses and your team's calendars side by side. List every recurring task where a human is paid to move information from one place to another, transfers, not decisions. Price each one honestly: hourly cost times hours per month, including your own time at your real rate. Rank by annual cost, then filter for tasks with a clear right answer, meaning a competent person would produce roughly the same output every time. Take the top line item and ask two questions: would you pay half that number for software that did it, and would ten other companies in your position pay the same. If both answers are yes, you have found a business. If only the first is yes, you have found an internal tool worth building anyway. The tell that you are onto something is embarrassment: if the task feels too mundane to mention on a podcast, that usually means nobody is building it.

Glossary

SOAP note

Subjective, Objective, Assessment, Plan. The standard structure of a clinical record. Veterinarians are legally required to keep one after every consultation.

AI scribe

Software that listens to a clinical encounter and drafts the record automatically.

Ambient recording

Capturing the consultation naturally without the clinician changing behavior, no dictation or commands required.

PIMS

Practice Information Management System. The clinic's system of record. CoVet integrates with Cornerstone, AVImark, Provet Cloud, ImproMed, Instinct, Shepherd, and others.

Model routing

Directing each task to the model best suited for it based on speed, cost, and quality rather than defaulting to the newest release.

Open source models

Publicly released model weights that a company can run on its own hardware, which resolves both cost and data residency concerns.

SOC 2 Type 2

An audited attestation that security controls operated effectively over a period of time, not just at a point in time.

HIPAA

United States health data regulation. Does not strictly govern veterinary records, which is exactly why CoVet's voluntary adherence is a differentiator.

PIPEDA

Canada's federal private sector privacy law.

TAM

Total addressable market. North America alone holds over one hundred thousand veterinarians, before adding equine, livestock, and international.

Seed strapping

Raising a small amount of outside capital and then funding growth from revenue instead of successive large rounds.

gTLD

Generic top level domain. The .vet extension launched in 2014 through ICANN's new gTLD program and is open to anyone.

Q&A: What Founders Ask After This Episode

What does CoVet do?

CoVet is an AI copilot for veterinarians. It records consultations through ambient capture on a phone, tablet, or clinic computer, then automatically drafts SOAP notes, discharge summaries, referral letters, and other required documentation, saving veterinary teams more than two hours a day according to the company.

Who founded CoVet and when?

CoVet was founded in late 2023 by Yannick Bloem, a former senior software engineer at Apple who serves as CEO and CTO, alongside Mike Parent, now COO, and Dr. Mike Mossop, a veterinarian who serves as Chief Veterinary Officer. The product launched to first customers in early 2024.

How do AI companies decide which model to use for which task?

CoVet maintains internal routing technology that assigns tasks to models based on speed, cost, and quality. An in-house team of over 30 veterinary professionals runs the evaluations, because domain experts rather than engineers are qualified to judge clinical output quality. Yannick's position is that the newest frontier model is frequently not the best choice for a specific vertical use case, since frontier models are heavily optimized for coding and general reasoning.

What compliance standards apply to veterinary AI software?

Veterinary data faces lighter regulation than human health data, and HIPAA does not strictly apply. CoVet voluntarily built to human healthcare standards, holding SOC 2 Type 2, ISO 27001, GDPR, HIPAA, and PIPEDA compliance, with contractual guarantees that customer data is not used to train models. The strategic reason is market access: adhering to the strictest standard lets the company sell into any geography and to enterprise veterinary groups without rebuilding.

Can you scale a startup to 130 employees without raising large venture rounds?

Yannick says CoVet did, funding most of its growth from customer cash flow. Early capital came from angel investors who were veterinarians that had sold clinics, adding domain validation alongside money. The company later closed a growth round with Off Leash Capital, the Chicago Pacific Founders Pet Fund, in September 2025.

Are niche industries good markets for AI startups?

Yes, per this episode. The veterinary market includes over one hundred thousand veterinarians in North America alone, plus equine, livestock, and international segments, and pet spending has grown consistently through economic cycles. Yannick argues the best AI opportunities today are unglamorous, expensive, manual workflows that companies already pay humans to perform, such as sales tax compliance and international paperwork.

How does CoVet acquire customers?

Two parallel motions: a bottoms up independent veterinarian motion with sales reps and a 14 day free trial requiring no credit card, and a top down enterprise motion selling to consolidating veterinary groups. The company also runs a student program giving free unlimited access throughout veterinary training plus 8 months after graduation, alongside school partnerships such as Colorado State University and state association member benefits.

URLs Mentioned in the Episode

Links & Resources