AI that helps the people who care for the animals
We are building AI into the records you already keep, to point the technician at the cage that needs a look and to take the paperwork off the veterinarian. None of it is live yet. Here is what is coming, how far along each piece is, and the rules all of it is built under.
What we are building
Each feature names its stage. Building means it exists and is being tested with a partner facility; In design means the workflow is drawn and reviewed with veterinarians; Exploring means we are still asking whether it belongs.
- Building
Photo welfare checks
On the mobile app, a technician snaps a photo of the animal at the cage. The app suggests what it may be seeing, such as hunched posture, a wound, barbering, a mass, or wet bedding, and files the photo with the case. The technician reports; the veterinarian decides.
- Building
Concierge chat
Ask your own records a question. Which cages on a protocol are over capacity, what an animal was given last week, how many animals a PI used this quarter. Every answer cites the record it came from.
- In design
Census by photo
Photograph the cage card and the cage. The headcount and the record are reconciled on the spot, and a mismatch becomes a finding for someone to sign off, never a silent correction.
- In design
Clinical note drafting
A technician’s observation becomes a structured case entry, with the signs named from your own library and the humane endpoints it comes near, ready for the veterinarian to review and sign.
- In design
Dose and plan assistant
Drafts the dose from the formulary and the animal’s weight, points out where an order differs from the default, and leaves the order for the veterinarian to approve.
- Building
Protocol number forecasting
Projects when a protocol will reach its approved number from intake and breeding trends, so the amendment is filed before the overage, not after.
- In design
IACUC and USDA report drafting
Semi-annual program review, annual USDA counts by pain category and post-approval monitoring summaries, drafted from the records you already keep and checked by a person before they leave the building.
- Exploring
Environmental anomaly alerts
Temperature, humidity and light-cycle logs watched for drift, with the room flagged before the next round rather than found at the end of the week.
How we build it
The same rules Vivocare already keeps for people apply to the model. These are not settings; they are how the features are designed.
The veterinarian decides
AI suggests, flags and drafts. It never signs. Every clinical decision stays with a named person who can see what was suggested and why.
Every suggestion is on the record
What the model proposed, what the person did with it and when, in the same audit trail as everything else, so an inspector can follow it.
Your records are yours
Your data is not used to train models, and nothing leaves your workspace without your say. The vendors we use are named in our data and security notes.
Help, don’t police
AI surfaces what needs a look and takes work off the desk. It does not block the round, refuse the dose or decide the finding. People do.
Design partners wanted
We build each of these with a working vivarium: a few rooms, a veterinarian who will tell us when we are wrong, and real rounds. If that could be you, we would like to talk.
See Vivocare with your own rooms and protocols.
A 45-minute walkthrough. We map one of your rooms and run a day of health checks, clinical cases and treatments on it, so you can judge the fit on your own work.