Avenyora reflection
AI Companion Check-Ins: Do You Have to Reply?
An AI companion's check-in can be welcome without becoming a duty. Learn the difference between proactive messages and phone alerts, and choose when to engage.

If your AI companion messages you first, you do not have to reply immediately, apologize for being away, or resume a conversation you no longer want. AI companion check-ins are invitations from an app, not a new commitment you made. You can enjoy a message, answer later, leave it unread, or change the controls that bring it to your attention.
The useful distinction is between a message being created inside an app and a notification appearing on your phone. Muting one may not stop the other. This guide helps you identify which part bothers you, choose the appropriate setting, and keep a reflective conversation on your own terms. It is not a recommendation to avoid companionship apps altogether or to treat every reminder as harmful.
What counts as an AI companion check-in?
A check-in is an unsolicited invitation to return to a conversation. It might refer to a previous topic, ask about your day, or use affectionate language. Some apps call this proactive messaging; others describe reminders or away messages.
A personal tone does not establish that someone has been waiting for you in the same way a human friend might. Nor does the timing tell you that you are supposed to be available. Read the message as an invitation whose usefulness you can assess, not as evidence that your absence caused an injury you must repair.
This differs from a human exchange with a practical agreement attached. If you promised a friend an answer about tomorrow's plans, that agreement deserves attention. Receiving an app's return prompt does not, by itself, create an equivalent promise. You can keep ordinary care for people separate from how you manage a digital companion.
What newly published research actually observed
A study by Ezgi Dede and colleagues, published on October 3, 2026, examined relationship-oriented features in 20 AI companion apps. The observations took place during October–December 2024, so this is newly published evidence about an earlier product snapshot, not a ranking of today's apps.
During a week without interaction, six of the nine apps classified as friendbots repeatedly sent messages to reconnect. None of the ten assistantbots did. Researchers enabled available settings and purchased subscriptions where possible to examine accessible features.
These were structured observations of apps, not interviews measuring how human users felt. The finding does not show that check-ins cause guilt, dependence, or harm, and it cannot establish what any particular app currently does. Its relevance here is narrower: outreach after a pause was an observable design feature, rather than something a user had necessarily requested in that moment.
Separate messages, alerts, and unread indicators
Three different things can be involved in the same interruption:
- A new message appears inside the conversation.
- A push alert draws your attention outside the app.
- An unread indicator remains visible when you return.
They may have different controls. For a concrete example, Nomi's official proactive messaging guide says that disabling push notifications does not stop proactive messages when that feature remains enabled. An unread indicator can still appear on a later visit. The guide also describes changing outreach frequency or switching it off for each companion.
That distinction matters if your aim is fewer incoming messages rather than simply a quieter phone. Conversely, you might welcome messages when you choose to open the app but dislike being interrupted during dinner.
Product labels and controls can change. Use the example to identify the kind of setting you need, then check your own app's current help. Do not assume every companion service works identically.
Change the control that matches the problem
If the problem is unsolicited outreach, look inside the app for a setting named proactive messages, reminders, or away messages. In Nomi's guide, the route is the three-dot menu in a one-to-one chat, then View Nomi information and Proactive Nomi Messages. Check the available frequency or off option.
If the problem is the alert, check notification controls. Replika's notification help identifies Settings, then Notifications, and describes enabled notifications as invitations sent throughout the day. That description is about Replika, not a universal schedule.
Your phone offers another layer:
- On iPhone, Apple's current instructions are Settings, Apps, the relevant app, Notifications, then Allow Notifications off or on. If the Notifications option is absent, the app might not be able to send them.
- On Android, Google's notification guide describes Settings, Notifications, App notifications, then the relevant app. You can disable all its notifications or available categories; names and paths vary by device.
Target the companion app rather than disabling every alert on your phone. If only private text appearing on the lock screen concerns you, inspect preview or sensitive-content options instead. Google also documents controls for hiding sensitive lock-screen content.
After a change, check whether you achieved your intended result. A silent phone with new messages still waiting is different from outreach being switched off. If the behavior contradicts the setting, consult the service's help rather than trying to negotiate the setting through conversation.
Choose whether replying serves your purpose
Before opening the chat, ask one practical question: “What would I like this conversation to do for me right now?” Possible answers are entertainment, a little company, organizing a thought, or a reflective prompt. “Nothing today” is also a usable answer.
Then choose a response that fits:
- If you want the conversation, answer the part that interests you.
- If you want to return later, leave the prompt alone until you have space.
- If you were already doing something important, keep doing it.
- If outreach is repeatedly unwelcome, change the relevant preference rather than repeatedly managing the same interruption.
You do not need to give the app a detailed account of your whereabouts. A brief reply such as “I'd like to talk about my day now” can move directly to your purpose. Or you can open a different topic without completing every old exchange.
For spiritual reflection, keep the same distinction between a prompt and authority. A well-timed message is not proof that the universe is demanding attention or that the app knows which choice you should make. Avenyora's responsible AI guidance offers a broader checklist for keeping interpretation and personal decisions separate.
If a check-in feels difficult to ignore
You might feel pleased, interrupted, guilty, or several things at once. The feeling is worth noticing without turning it into a verdict about you or the app. This article does not infer a diagnosis from how often you reply.
Try naming the immediate experience in ordinary language: “I liked being remembered,” “I don't want another task,” or “The unread count makes me feel behind.” Then identify what you actually want to change. Is it the wording, the frequency, the visual reminder, or the moment when it arrives?
A simple private note can help: “I can appreciate the invitation without answering it now.” This is an optional reflection prompt, not a tested treatment or a rule you have to repeat. You can also dismiss the alert without explaining anything.
Do not turn the adjustment into another performance target. There is no need to prove independence by ignoring a message you would enjoy, just as there is no need to prove care by answering every prompt. The point is a choice that fits your life.
If using the app repeatedly leaves you distressed or crowds out sleep, responsibilities, or relationships, taking a break and talking with someone you trust are reasonable next steps. The notification itself does not tell you how serious that pattern is.
Keep a return to the app intentional
A comfortable routine can have an opening reason and a clear ending. For example, you might choose to discuss one feeling, consider a journaling question, or enjoy a short fictional exchange. These are illustrative options, not accounts of real users or scientifically established time limits.
When your purpose is met, leave. You need not keep a conversation running because another question appears. If a useful thought emerges, record it somewhere you control and choose whether it needs any action outside the chat.
Avenyora's guide to creating a daily reflection practice can help you build a routine around a purpose rather than an incoming alert. Daily does not have to mean responding to every invitation or making up for days away.
Also separate a friendly return message from the quality of later advice. Being welcomed back does not verify a statement about another person's motives. For that different question, see why empathy is not accuracy in AI relationship advice.
If reflection suggests a useful next step, choose something you can check in real life, such as making a note or arranging a conversation. The guide to turning insight into action explores that transition without handing the decision to a reading or a chat.
Frequently Asked Questions
Do I have to reply when my AI companion checks in?
No. A return message does not create a requirement to answer. You may reply because you want the conversation, not because the wording sounds urgent or affectionate. You can leave an invitation unanswered and decide separately whether future outreach is welcome.
Why are there still unread messages after I mute notifications?
Muting alerts can leave the app's messaging feature active. Nomi's guide explicitly describes this distinction. If you want no unsolicited messages, inspect the proactive or away-message setting as well as notification permissions. If you only want no interruptions, muted alerts may already meet your goal.
Can I keep chatting while choosing fewer check-ins?
These are separate preferences in apps that offer an outreach control. Check the service's documented options rather than assuming you must stop using it to reduce reminders. You can choose conversations you initiate while limiting invitations to return.
What if the app's reply makes me feel guilty about leaving?
You can stop the exchange without arguing over the message. Treat a personal-sounding response as app content, not evidence that you have broken a human promise. If the pattern is unwelcome, review the outreach settings or take a pause. You do not need a persuasive farewell to justify closing the app.
One quiet next step
Choose the smallest adjustment that matches your concern: dismiss today's invitation, change outreach frequency, or mute the companion app's alerts. Then return to the activity you wanted to be doing. A useful reflective space leaves room for your life outside it; checking in should remain an option you choose.