The most important parts of caring for a pet at the end-of-life will never be automated. A lot of what surrounds it can be, and probably should be.
If you work in veterinary care, you will have noticed that everyone has started talking about AI. Some of it is excitement, some of it is worry, and a surprising amount of it is people quietly admitting they aren’t sure what is real.
I look after IT and systems at Dignipets, where we provide in-home end-of-life care for pets. This week I have set aside time to look properly at what AI means for a business like ours, and for the wider profession. This post is where I’ve got to. It’s my opinion rather than a forecast, and I expect to change parts of it.
What we can use today
Most of what is useful right now is unglamorous. It takes effort out of the work around the care, so there is more time for the care itself.
- Reading long documents so people don’t have to. We have used AI to turn lengthy regulatory documents, including those published by the CMA, into plain-English summaries to work from. A person still checks them against the source, but it saves a lot of reading.
- Finding things quickly. We are building AI-assisted search into our client management system, so the team can ask a question in everyday words and get to the right answer, rather than clicking through screens.
- First drafts. Internal notes, procedures and newsletters are quicker to start. Nothing goes out until a person has rewritten it in our own voice.
- Meeting notes. Recorded and summarised, so actions don’t get lost.
Across the profession the pattern is similar: tools that write up a consultation, phone systems that handle routine questions out of hours, and reminders and follow-ups that send themselves. All of it saves time. None of it is the care.
What is coming next
Some of this is already happening, and given how fast things are moving I expect the rest within months rather than years. Three things stand out.
- AI is already inside the software we use. Phones, email, diaries and customer systems have all added AI features, whether we asked for them or not, and more arrive with every update. That makes it more important to know what each one does with our information.
- Voice will get better. AI that answers the phone already exists and will improve steadily at routine calls. A call from someone whose pet is dying is not a routine call, so how we use it will matter far more than whether we can.
- Planning and follow-up will get smarter. Scheduling visits and routes for mobile teams, noticing when a family may need a check-in, and tracking quality-of-life information over time are all jobs software is well placed to help with.
These are informed expectations, not promises. This area is changing almost weekly.
Where it could end up
Looking further ahead, I can imagine most of the administration around a visit being handled quietly in the background: records written, letters prepared, follow-ups sensibly timed, so that a vet’s whole attention is on the animal and the family in front of them. That would be a real gain, and it is the future I’d like us to build towards.
There is another direction too. Tools already exist that generate messages “from” a pet that has died, or chatbots modelled on a pet’s personality. Some people may find that comforting. Others may find it unsettling, or it may complicate their grief. It is a reminder that something being possible doesn’t mean it should be offered at the most vulnerable moment of someone’s life. As a profession, I think we should have a view on that before it is offered to our clients, not after.
What AI shouldn’t replace
This is the part I feel most strongly about.
- The person in the room. Families remember who was there: how they spoke, how they handled the animal, when they chose to say nothing. That presence is the service.
- The judgement. Deciding with a family when the time is right is clinical, ethical and personal all at once. AI can help organise information such as quality-of-life scores. It should never make that decision, or nudge it.
- The first conversation. Someone ringing to arrange a goodbye should reach a person who understands what the call means, not a script.
- Sympathy in our own words. A condolence message that a machine wrote and nobody read is worse than silence.
- Accountability. When something goes wrong there must be a named person who answers for it. “The AI did it” is never an acceptable answer to a grieving family.
How we are approaching it
Calmly and practically.
- Start small and check everything. We try tools on real work, and a person reviews what comes out.
- Protect people’s information. We hold some of the most personal details there are. Free AI accounts can keep what you type and even use it to train their models, so we are setting clear rules: approved tools on company accounts under business terms, and nothing about clients, money or staff going into anything else.
- Keep humans responsible for anything a client sees.
- Choose tools we can swap. Nobody knows which tools will still be around in two years, so we would rather stay flexible than commit to something we can’t leave.
If you are a vet, nurse or practice owner thinking about the same questions, I would genuinely like to hear how you are approaching them.
How this post was written
This post was written with the help of AI. I used Claude to help draft and structure it, then read, edited and rewrote it myself. The opinions are mine, and so is the responsibility for what it says. That is the same standard I’d ask of anyone using these tools.
