All use cases

Prevent / Care

Cycle-anomaly health flag with safe escalation

She has logged faithfully for months, and her cycles have quietly drifted out of normal range — long enough that it is worth someone looking, but she has no reason to notice on her own.

Get aheadRisk triggerBehavioral triggerPushChatIn-appWellness (B2C apps)Health & Telehealth

What it does

A period tracker holds a signal a user rarely reads herself: when her own logged cycle or period length has been repeatedly outside the typical range. Left unsurfaced, a pattern that could point to stress, PCOS, thyroid issues, or endometriosis sits in the data for months while she assumes everything is fine. The agent watches the trend rather than a single odd month, raises it gently when it persists, gives plain education on what can cause it, and routes her to a symptom checker or clinician — early care without ever putting a diagnosis on it.

How it works

  1. 1
    Trigger. the user's logged cycle length or period length falls outside the normal range repeatedly across recent cycles, detected against her own history.
  2. 2
    Decision. the agent confirms this is a sustained pattern, not one irregular month, and checks logging completeness so it is not flagging a gap in data; it weighs how to frame it without alarming her.
  3. 3
    Action. a gentle, non-diagnostic message names the pattern, offers context on common causes, and points to the in-app symptom checker or a clinician route — framed as worth checking, not as a verdict.
  4. 4
    Follow-through. if she engages, the agent hands a clean summary of the logged pattern to the care pathway; if she dismisses it, it backs off and only re-raises if the trend worsens, never nagging.
Example
AgentI want to flag something gently — your last few cycles have run noticeably longer than your own usual range. That can have lots of ordinary explanations like stress, and it is not a diagnosis, but a pattern like this is worth a professional eye. Want me to open the symptom check or help you find a clinician?
UserHuh, I hadn't clocked that. Is it serious?
AgentMost often it is something manageable, and I genuinely can't tell you what it is — that is for a clinician. What I can do is share the dates and lengths you've logged so you don't have to recount them. Shall I set that up?

Configuration

How the agent is wired for this use case.

Triggerrepeated out-of-range cycle-length or period-length values in the user's own logged history, evaluated by the app backend against her personal baseline
Tools & actions
  • App backend (cycle log) · read logged cycle and period lengths and compute deviation from the user's own baseline range
  • Knowledge base · pull plain-language, non-diagnostic education on common causes of cycle irregularity
  • Messaging channel · deliver the gentle flag with a symptom-checker or clinician route
  • Scheduling / care pathway · open the in-app symptom checker or hand a clean summary of the logged pattern to a clinician route
  • Memory store · record that the flag was raised, the user's response, and suppress repeat nudges unless the trend worsens
Autonomyanomaly detection and the educational flag run unattended under a SaMD-safe content policy with every message judge-gated; the agent never names a condition, never gives a diagnosis, and never recommends a medication — it detects and routes only
Channelspush · chat · in-app
Escalationa sustained anomaly, or any acute symptom the user adds in reply (severe pain, heavy abnormal bleeding), routes to a clinician or symptom-checker pathway with the logged pattern attached

What you need

The inputs this use case runs on. Your channels stay yours; the agent supplies the judgment.

Signals

cycle-start and period-length log events, logging-completeness state

Data

the user's own cycle history and personal baseline range, non-diagnostic education content, consent to receive health flags, persistent memory of prior flags

Guardrails

SaMD-safe posture — detect and route, never diagnose, never name a condition, never advise a medication; gentle non-alarming framing; suppress repeat nudges on dismissal; acute symptoms in reply escalate immediately; judge gating and audit trail on every flag

Metrics it moves

  • safe-escalation-rateup, as quantitative anomalies in self-logged data reliably reach a clinician or symptom checker instead of sitting unread
  • contact-ratedown, because users get an early, self-serve route to context rather than arriving anxious through support
  • csatup, since the product feels like it is looking out for her health, not just storing her data

Related use cases

Prevent / Care

Clinical red-flag detection and safe escalation

the chat-side red-flag sibling; this card fires on quantitative logged anomalies

Prevent / Care

Prediction why-isnt-it-showing deflect and escalate

the profile-specific escalation route in the same app

Engage / Lifecycle

Hormone-result interpretation grounded in trend history

the engagement layer reading the same cycle data over time

See it on your own customer journey

Bring one drop-off, one churn cliff, or one silent segment. We will show you what a proactive agent with memory and judgment does with it.

Book a demo