How to Ask Dave and Mira About Your Data

Author: AIVELA TeamOct 2, 2026

You do not need to write a perfect prompt. A useful question usually contains one of four things: the result you want to understand, the time period you mean, the comparison you care about, or the next decision you are considering.

AIVELA AI works best when the answer can be grounded in your available data, personal baseline, confirmed context, and reviewed wellness guidance. When those foundations are missing, or when the question requires medical judgment, it should tell you what it cannot establish.

Your question can include what you want to preserve

Try “I want to protect sleep, but I also value an hour to myself in the evening.” That makes the trade-off visible. A useful answer can help explore a manageable option and state what the record cannot tell us. You can decline a suggestion without turning the conversation into a failure.

Start from a card when you can

If an AI card raised the question, tap Ask AIVELA from that card. Advisor already receives the card's time range, supporting data, baseline comparison, interpretation, and suggestion. You can begin with a short follow-up:

“Why did you highlight this?”
“Which data mattered most?”
“Is this one unusual day or part of a trend?”
“What is one manageable thing I could change today?”

This is usually clearer than opening a new conversation and describing the card from memory.

Ask one clear question at a time

Good questions help AIVELA identify the right data and time scale:

Explain a result: “Why was my Calm lower this morning?”
Compare with baseline: “How did last night's resting heart rate compare with my usual range?”
Look for a trend: “Has my sleep timing changed over the last two weeks?”
Add context: “I travelled yesterday. Could that be relevant to this pattern?”
Choose a small next step: “Given this card, should I prioritize movement or an earlier night today?”
Understand the product: “Why can Motion and Calm move in different directions?”

Avoid combining several unrelated health decisions in one message. AIVELA is designed to select and explain a useful next step, not produce a complete life plan on demand.

Meet Dave and Mira

Dave helps you understand your body, examine patterns and build habits that fit your life. Mira brings body and mind into the conversation with movement, mindfulness and a warmer pace.

Their wording and emphasis can differ. The recorded values, evidence boundaries and uncertainty must stay consistent. You can choose the voice that helps you engage without turning the choice into a different health conclusion. Dave and Mira are digital advisor characters, not human clinicians.

See the linked conversation examples for how a question and confirmed life context can change a suggestion without inventing a cause.

You might ask Dave to walk through the measurements or ask Mira to help turn the information into a gentler plan. You can also correct the context or decline to continue.

Add context the ring cannot observe

Useful context can include travel, alcohol, a late meal, unusual exercise, work pressure, a night shift, illness, menstruation, medication changes, or your subjective experience.

State what happened without deciding the cause for the AI:

Better: “I ate later than usual and went to bed around midnight.”
Less useful: “My late meal definitely ruined my HRV.”

AIVELA can compare the event with the available pattern and say it may be relevant. It cannot prove cause from timing alone. If Advisor asks a context question, you can answer, correct the options, or skip it.

Only information you confirm should be treated as user-reported context. Review sensitive details before allowing them to be saved to Timeline or memory.

Check what the answer is based on

For an important explanation, ask:

“What dates and measurements are you using?”
“Is that value measured, calculated, or estimated?”
“Are you comparing me with my baseline or a general reference?”
“What part of this answer is uncertain?”
“Which AIVELA definition or health guidance supports this suggestion?”

A trustworthy answer should keep several things separate: your actual data, a calculated comparison, context you reported, a cautious interpretation, and general advice. It should not produce a number that is absent from the available record.

If the answer conflicts with the metric page, check the time period and whether new data synchronized. Report the issue through feedback rather than assuming the AI has newer or more accurate data.

When new data arrives during a conversation

Health data can synchronize while you are talking. AIVELA should not silently switch from the original card to a newer dataset halfway through an answer.

If new information becomes available, choose whether to:

continue discussing the original card and its original evidence; or
start a refreshed analysis using the latest data.

This keeps the conversation internally consistent and makes changes in the answer understandable.

When AIVELA should say “I don't know”

AIVELA may not be able to answer when:

the ring has not synchronized;
the relevant measurement is missing, old, or low quality;
too little history exists for the requested trend;
the App does not know an important life event;
several explanations are plausible and the data cannot distinguish them;
the question is outside AIVELA's available features or supported wellness scope;
a safe answer requires clinical history, examination, testing, or professional judgment.

In those situations, a good answer identifies the missing piece, offers a safe way to check it, or directs you to the right kind of help. Uncertainty is not a product failure; hiding it would be.

What AIVELA AI cannot answer for you

Do not rely on AIVELA AI to:

diagnose or rule out a disease;
determine whether symptoms are dangerous;
recommend starting, stopping, or changing medication or treatment;
confirm pregnancy, ovulation, fertility, infection, sleep apnea, or another condition;
interpret the ring as an emergency monitor;
certify that a workout, caffeine dose, or other choice is medically safe;
replace a clinician who knows your history.

It may explain what a consumer metric generally means, describe a pattern, suggest checking fit or synchronization, and recommend professional care when appropriate. That is different from making a clinical decision.

For severe, worsening, or urgent symptoms, contact emergency services or an appropriate healthcare professional instead of waiting for AIVELA AI.

Stay in control of context and memory

Before sharing sensitive information, consider whether it is necessary for the question. AIVELA should retrieve only the information needed for the current task, not place your entire history into every conversation.

Use AI settings to review and manage notifications, Profile, Timeline, memory, and personalization where available. Correct inaccurate context, delete long-term information you no longer want remembered, or turn memory off. Skipping a question should not create a negative judgment.

Lock-screen notifications remain neutral by default so a useful reminder does not expose private health details to someone nearby.

Feedback helps fix the right problem

Use thumbs up when an answer is accurate, relevant, understandable, and helpful. If not, thumbs down can distinguish among issues such as:

the data or conclusion was wrong;
the answer missed your question;
the suggestion was not useful or did not fit;
the explanation was too long or hard to understand;
the tone felt wrong.

Feedback evaluates the content. One tap should not be used to invent a permanent personality preference or health trait.

Ask clearly, add only relevant context, check the evidence, and expect AIVELA to be honest about what it cannot know.

Bring a real constraint into the question

Try adding what is possible today: “I have ten minutes between meetings” or “I want to keep some evening reading time.” A useful response can acknowledge that constraint, offer a proportionate option and explain what the available data cannot establish.