My mother was a hospice nurse for twenty years, and I grew up hearing about her work. Families at the end of a life are exhausted, frightened and very good at noticing when someone is selling them something. I think about them every time I write a sentence on this site.

Senior care is full of big promises. Apps that "improve memory." Programs that "reduce agitation by 40 percent." AI that "understands" your mother. Most of it is not true, and the people who work in this field can smell it from across the room. So this page is the opposite of a feature list. It is what Porchlight does not do, and why. If you are deciding whether to trust us with someone you love, this is the page I would want you to read first.

It is not therapy, and it does not treat anything

There is real research on reminiscence and life review, and we cite it often. Structured life review has been linked to lower depressive symptoms in older adults, and reminiscence in dementia shows small benefits in some settings.12 But almost all of that research involved trained facilitators running sessions, often with people who were already depressed. Nobody has run a trial of a self-serve life story app, ours included.

So Porchlight does not claim to improve memory, slow dementia, reduce depression or loneliness, or calm agitation. It is a conversation and keepsake tool. It is not a medical device and it is not a substitute for clinical care. When you see us cite research, we are describing what the field has found, not making promises about our product.

AI does not choose the questions

Every question a resident hears was written by a person, and they come in a fixed order. The sequence starts gentle (childhood, home, food, music) and moves toward bigger chapters later. Residents can skip anything.

We could have let an AI improvise follow-up questions. We chose not to. A fixed bank means staff and families can see exactly what will be asked. It means a question about a late spouse does not appear out of nowhere because a model guessed wrong. It means nothing surprising happens to a vulnerable person at the end of a recording session when nobody is in the room. Predictable is a feature here.

Predictable is a feature when the person answering is vulnerable.

AI does not diagnose, and it does not detect distress

Porchlight does not listen for signs of decline, depression or distress, and it does not flag them. I want to say that plainly, because some products in this space imply they can. Telling the difference between a hard memory told with feeling and a person in real trouble is a human job.

That is why the design leans on people. Staff can mark topics to avoid on a resident's profile. Any question can be skipped. Each resident has a stage setting, and in the later stage a session stops after two unanswered questions and leaves a note for staff. The research is clear that some memories hurt: rehearsing old grievances is linked to worse mood, and for some veterans, reminiscing can reawaken wartime trauma.34 The answer to that is not a detector. It is a skip button, a gentle bank, and a person who checks in.

AI does not write the person's story for them

AI does two narrow jobs in Porchlight. It transcribes what the resident said, and it tidies the transcript lightly so it reads cleanly, removing the "ums" and false starts. A check then compares the tidy version against what was actually said, to catch anything added. The original audio is always kept, and staff and family can listen to it.

It does not invent details, summarize a life into a biography, or rewrite someone's words into prettier ones. Your father's voice saying it his way is the whole point.

No voice cloning, no chatbot grandparent

Porchlight will never synthesize a resident's voice or a family member's voice. There is no "talk to Grandma after she's gone" feature and there will not be one. When a grandchild sends a voice note, it is the grandchild. When a resident tells a story, it is the resident. We think a recording is precious exactly because it is real.

It does not rush anyone

There is no time limit on an answer. After a resident finishes a story and goes quiet for a few seconds, a short countdown appears before the next question; if they start talking again, it cancels. Staff can switch that off for anyone who would rather tap to move on, and in the later-stage setting nothing moves on by itself at all. A pause is often a person thinking, and we would rather wait than interrupt.

It does not replace your assessment, your care plan, or your compliance work

Nursing home rules require a resident's life history as part of the assessment, and many state assisted living rules ask for preferences and routines.5 Porchlight can help keep that part of the file current and get it in front of staff. It does not make anyone compliant, and it is not an electronic health record, an MDS tool or a service plan. Your clinical team still owns all of that. (We wrote about what the rules actually say.)

It does not replace people

The interventions that have actually reduced conflict between families and staff were human ones: training, meetings, agreements.6 A tablet cannot do that work. Porchlight gives people something to talk about. The aide still has to walk in and say, "I heard you used to fish the Snake River." The daughter still has to visit. The activity director is still the heart of it. If a community buys Porchlight to replace any of those people, it will not work.

It does not make anyone participate

Not everyone wants their stories kept. In a large US poll, 31 percent of adults said they would not want their conversations preserved.7 That is a legitimate choice. Residents can decline, stop, or skip. Communities use a paper consent form with an opt-out, and there is a plain-language "who sees what" page explaining where stories go. For residents who would rather not look back at all, there is a "Just for fun" track of present-tense questions, like settling a friendly argument.

It does not ask you to trust the cloud forever

Online services get bought, change their plans and shut down. One researcher who interviewed bereaved people found they took real comfort in saved voicemails and messages, and lived with a fear of a "second loss" if the device or platform disappeared.8 So families can download every recording and keep their own copies, and can order a printed memory book. Paper outlasts any company, including ours.

What we cannot claim yet

A few things we will not say until they are true:

What it does do

After all that, here is the short version of what is left. A resident taps one button and tells their story to questions read aloud. Their family hears it, keeps it, and answers with their own voice. The aide on the next shift knows one true thing about them. That is not a miracle. It is just something that should happen for every person in care and almost never does.

See exactly what a resident hears

The live demo shows the real question bank, the staff briefing and the family page, with sample residents.

Explore a live demo Book a demo

Sources & notes

  1. Westerhof GJ, Slatman S. In search of the best evidence for life review therapy to reduce depressive symptoms in older adults: a meta-analysis of randomized controlled trials. Clin Psychol Sci Pract 2019;26(4):e12301. doi:10.1111/cpsp.12301.
  2. Woods B, et al. Reminiscence therapy for dementia. Cochrane Database of Systematic Reviews 2018;3:CD001120. doi:10.1002/14651858.CD001120.pub3.
  3. O'Rourke N, Cappeliez P, Claxton A. Aging Ment Health 2011;15(2):272-281 PubMed; Korte J, Bohlmeijer ET, Westerhof GJ, Pot AM. Aging Ment Health 2011;15(5):638-646 PubMed. Both link "bitterness revival" reminiscence with more depressive symptoms.
  4. Davison EH, et al. Later-adulthood trauma reengagement in aging combat veterans. Gerontologist 2016;56(1):14-21. PubMed.
  5. 42 CFR 483.20(b)(1), eCFR. ecfr.gov; Porchlight, state assisted living rules.
  6. Pillemer K, et al. Partners in Caregiving randomized trial in 20 nursing homes. Gerontologist 2003;43 Spec No 2:96-106. PubMed.
  7. YouGov, survey of 6,090 US adults, May 31 to June 1, 2022. yougov.com.
  8. Bassett DJ, on the bereaved and digital messages, writing in The Conversation. theconversation.com.

Scenarios marked as illustrative are not real named individuals. Porchlight does not fabricate customers, testimonials, or outcome metrics.