We changed the first digit of Animana. Here’s what it cost.

4 mins

It’s Tuesday, just after nine. The waiting room is full. There’s a dog that shouldn’t be next to a cat, a pet owner who has already asked twice how much longer, and a nurse trying to find out whether the patient in front of her is allergic to anything.

And nothing loads.

That morning is the only moment anyone thinks about the foundations their software is built on. Not the interface. Not the feature you saw in a release note. The layer nobody talks about: the framework, the security patches, the performance headroom, the parts that either hold or don’t.

They usually hold. That’s the problem. Foundations don’t warn you gradually. They hold, and hold, and then one morning they don’t.

Which brings us to a number.

For years, IDEXX Animana sat in the 8s. Most weeks brought a release: a fix, a refinement, or a field that finally behaves the way you expected it to the first time. 8.18, 8.19, 8.20. Quiet, incremental, forgettable: exactly what you want from the software running your practice.

It’s now 9.0.

That number wasn’t chosen. It was arrived at. Compare it to Apple, which last year skipped from iOS 18 to iOS 26 so the number would track the model year rather than the release count. Most version numbers work that way now. They’re labels, telling you when rather than what. This one moved because counting on in the 8s would have been misleading.

A large share of 9.0 went into precisely the layer you’re only supposed to notice on the morning it fails. Replacing that while practices are working on top of it is slow, deliberate work, and there is no honest shortcut. What it buys you is software that stays standing when it matters, keeps pace as your practice grows, and is ready for what’s coming, because the next few years will ask far more of that foundation than the last ten did.

It’s also why the changes you can see arrived the way they did.

A patient file built for what comes next

The clearest of those is the patient file.

It is being rebuilt in phases rather than swapped over one weekend, spread across 2025 and 2026, specifically so you can keep working while it happens. The new view is structured so you can take in a patient’s history at a glance instead of assembling it from fragments.

The bigger reason for the rebuild is harder to see from the screen. The old structure had reached its limits. A number of things practices have asked for over the years weren’t difficult to design, they were difficult to fit. The new structure removes that ceiling, and most of what it makes possible is still ahead of us. That’s rather the point of doing it now.

Weekly, not quarterly

A major version arrives every few years. Everything in between arrives constantly.

Recently that has included automatic appointment reminders you can refine by species, so a feline reminder doesn’t read like a canine one. Reminder history you can check from the appointment list without leaving the agenda. Bulk invoice download, up to a hundred at a time. Stock list work shaped by user feedback, including the ability to turn off quantity tracking on disposables so your order alerts stop crying wolf. Single sign-on across linked IDEXX solutions. More insurers supported for electronic claims.

Some of it is less a feature than a removal. Ordering IDEXX in-house tests no longer needs Interlink or WebAgent, and the invoicing follows automatically. Nothing was added to your screen. A step came out of your day that you had probably stopped noticing was there.

The clearest test of all of it came from outside. When the CMA changed what UK practices are required to show pet owners, price estimates were reworked to suit: quicker to create, easier to get at, and equipped with save options that return you to the patient file where you were already working. A regulator changed the rules, and the software changed with them.

All of it is published on the roadmap and releases page, along with what’s currently in research and in development. It’s public for a reason. You should be able to check whether we’re doing what we say we are.

About the promises

If you’ve been looking at what else is available, you’ll have noticed there’s a lot of it, and that it arrives sounding confident. Some of those claims will turn out to be true.

The question worth asking is what it costs to make them.

Announcing a feature is quick. Making it survive contact with a real practice is not. A busy Friday, an incomplete record, an unusual species, a batch credited after it was deleted, a caretaker who shouldn’t receive the reminder meant for the owner, and a regulator with a firm opinion about how you present estimates. That is most of the work, and it’s the part that never appears in a demonstration.

Animana and ezyVet have both been running real practices for years, and that has a specific value which isn’t nostalgia. Every situation in that list was met the hard way: reported by a practice, understood, and fixed. None of it was designed in advance, because none of it could have been. Newer software hasn’t had the chance to be corrected like that yet, and it will be your practice that finds out where the gaps are.

What that looks like when it works

The AI-assisted Consult has been in open beta with UK practices since the spring, free of charge while it stays in beta.

It works by voice. You start an AI-assisted consult and talk normally through the appointment, pauses, interruptions and background noise included. Animana then turns the recording into a structured consult note under History, Findings, Diagnosis, and Plan, picks out a weight if one was spoken, and files the transcript separately in the patient file. All it needs is a microphone, and an admin to switch it on under General Settings and the Integrations Hub.

The important part is what it doesn’t do. Nothing reaches the patient file until a clinician has read it and approved it, and it makes no attempt to diagnose anything. It reduces the typing and leaves the judgement where it belongs.

The word beta is doing real work in all of that. It means we want the feature used in real practices before we draw conclusions from it, and that what you report back decides what happens next. We’ve written separately about why we’d rather get it right than get it first, and about which veterinary uses of AI look safe today and which don’t.

What happens after 9.0

More releases, most weeks, the same as before.

If you’d like a say in what those contain, the research group is open, and the patient file is a fair illustration of what that input changes. The version number will go back to ticking along quietly for a while now. That’s usually a good sign.