Most of what is written about voice companions in dementia care assumes the caregiver’s own voice is the obvious goal, and that anyone choosing otherwise is settling. That is not what we see. Some families arrive certain they want their own voice and are right. Others arrive uneasy about it, and their unease is usually pointing at something real.
This is a straight look at when each one fits.
What is an assistant voice?
It is a warm, natural companion voice that is ready immediately. Nothing is recorded from you and nothing is recorded from your loved one, ever. For families curious about how personal voice capture works when they do choose that path, here is how that process works and what it involves.
What an assistant voice is not is a robocall or a smart speaker. The voice is only the surface. Underneath, an assistant voice runs on exactly the same foundation as a caregiver’s own voice does: the knowledge base you build about your loved one, their people, their routines, the places and stories that made them. It follows validation therapy, meeting your loved one inside their reality rather than pulling them out of it. It never corrects. It never says "you already asked me that."
It answers the calls your loved one makes, and it makes the check-in calls and gentle reminders you schedule. Everything works the same way. Only the timbre is different. The full 17-standard voice model applies in both cases.
Six reasons families choose the assistant voice
1. Hearing your voice sends them looking for you
This is the one families do not anticipate, and it is the most important.
For some people with dementia, particularly those with strong separation anxiety, hearing the voice of the person they are attached to activates the search rather than settling it. Where are you. Why aren’t you here. When are you coming. The voice becomes evidence that you exist somewhere nearby and are not present, which is precisely the thought that was causing the distress.
A warm voice that is not yours can offer the same comfort without opening that particular door. Plenty of people with dementia are settled beautifully by a kind stranger, and are undone by the person they miss most.
If your loved one’s calls are mostly a search for you specifically, this is worth trying.
2. The relationship is complicated
Not every family is a warm one. Some caregivers are looking after a parent who was harsh with them, or absent, or worse. Some are looking after a spouse in a marriage that had already grown difficult before the diagnosis.
There is a version of this where a daughter’s voice, delivered gently on a schedule, is a genuine gift to both of them. There is another version where being asked to record hours of warmth she does not feel is its own small cruelty, and where the voice itself carries a history the call does not need.
If putting your voice on those calls feels wrong, that feeling is worth respecting rather than overriding. The assistant voice exists so that difficult family histories do not disqualify anyone from getting help.
3. You are not the only one caring for them
When four siblings share the load, whose voice is it?
Choosing one turns a practical question into a status question, and families have enough of those already. A shared assistant voice sidesteps it entirely. Everyone contributes to the knowledge base, and nobody has to be the voice.
4. Recording your voice does not sit right with you
Some families feel uneasy about a companion using their own voice, and that is a reasonable feeling, not a hurdle to be talked out of. It is not the same worry for everyone. For some it is about whether it counts as deception. For others it is about hearing themselves and finding it uncanny. For others it is simply private.
We would rather a family use the assistant voice and stay comfortable than use their own voice and feel a knot about it every time the phone rings. The ethics of both choices are examined more closely in our piece on AI and ethical dementia care.
5. There is no time and no quiet right now
Setting up your own voice takes recording sessions in a quiet room. If you are in the middle of a hospital admission, a move into care, a bad month at work, or a house with three children in it, that is not a small ask.
The assistant voice is ready now. Some families start there, get the calls running this week, and set up their own voice later when things have settled. Switching is a setting, not a rebuild.
6. Someone else’s voice is the right voice, and you cannot record it
Occasionally the person your loved one is really searching for is a spouse who has died, or a sibling who is gone. That voice is not available, and recreating it from old recordings is not something we do. A warm, neutral companion voice is the honest answer here, and it carries none of the weight that an impersonation would.
Does an unfamiliar voice still work?
Yes, and it is worth being precise about why, because the honest answer has two halves.
The research base for simulated presence therapy, including the 2024 randomized controlled trial by Duan, Liu and Zhang in the International Journal of Neuroscience (PubMed 38646703), specifically involves a familiar voice. We are not going to claim that evidence for an assistant voice. It is not the same intervention.
What the assistant voice draws on is a different and well-established set of mechanisms. Warmth and calm tone settle people, which is the core of the Alzheimer Society of Canada’s communication guidance. Specific, personal knowledge does the heavy lifting: a voice that knows your mother grew up on a farm outside Peterborough, that her sister is Anne, that she worries about the dog, is doing something no generic voice can do regardless of timbre. And validation, meeting someone where they are rather than correcting them, works no matter who is speaking.
There is also a quieter effect. Consistency creates familiarity. A voice heard every day at the same time, always kind, always knowing who you are, becomes a known voice over weeks. It is not the same as a daughter’s voice. It is not nothing either.
If your loved one is deep into later-stage dementia and recognition of familiar voices is one of the few things still holding, your own voice is likely the stronger choice. In most other situations, the difference is smaller than families expect.
Side by side
| Your own voice | Assistant voice | |
|---|---|---|
| Setup | Recording sessions with Sarah | Ready right away, no recording |
| Recognition | Draws on voice recognition that often persists in dementia | Becomes familiar through consistent daily contact |
| Best when | Strong bond, later-stage dementia, your voice reliably settles them | Your voice triggers searching, complicated history, shared caregiving, or you would simply rather not |
| Knowledge of your loved one | Full knowledge base | Full knowledge base, identical |
| Answers their calls | Yes | Yes |
| Makes check-in calls and reminders | Yes | Yes |
| Your loved one’s voice recorded | Never | Never |
| Switchable later | Yes | Yes |
What stays the same, whichever you choose
- The knowledge base you build about your loved one, and everything the companion knows.
- Validation therapy, and never correcting or quizzing them.
- Honesty if they ask. If your loved one asks whether they are talking to a real person or a machine, they get a truthful answer, gently. Comfort is never built on a lie that will be found out. The same principle applies to therapeutic fibbing more broadly.
- Never saying that someone is on their way when they are not.
- Quiet hours, and reaching out being off until you turn it on.
- Your loved one’s own wishes taking priority. If they ask the calls to ease off, that is honored right away.
- Your loved one’s voice is never recorded or modeled, on any call, either way. How we protect them covers this in full.
Can I switch later?
Yes, at any time, and in both directions. A number of families start with the assistant voice because it is ready today, then set up their own voice a few weeks later once the calls are already part of the routine and the pressure has come off. Others do the reverse when the relationship or the response changes.
You can also run a fair test. Try one for a week or two, watch how your loved one is in the hours after a call rather than during it, and switch if the other one suits them better. That is a more reliable answer than anything anyone can tell you in advance.
So which should you choose?
Choose your own voice if the bond is warm, if your voice reliably settles them, or if dementia is advanced and familiarity is one of the last things holding.
Choose the assistant voice if hearing you sends them searching for you, if the relationship carries something you do not want on those calls, if several of you share the caregiving, if recording feels wrong, or if you simply need this working this week.
Neither is a lesser choice, and neither is the compromise. The thing that makes the call land is not only the voice. It is that someone on the other end knows who your loved one is, speaks warmly, and never makes them feel caught out.
"Almost every other option in this space works on one side of the problem: quieting the caregiver’s phone, while the person on the other end is left alone with the fear. KindredMind was built for both sides at once, and that holds true in either voice."
KindredMind doesn’t replace the caregiver. It’s built so the caregiver can be there for everyone.
Hear what a call sounds like.
Both voices, the same warmth. See how it works and start your free 14-day trial.
See how it works