All use cases

Prevent / Care

Overtraining and strain-creep early warning

Your recovery has been dropping for three mornings while you keep pushing hard — the agent sees the trend before you feel it and tells you to plan a lighter week.

Get aheadRisk triggerBehavioral triggerPushChatIn-appWellness (B2C apps)

What it does

Committed members trust a wearable to tell them when they are overdoing it, but the warning signs live in a multi-day trend that no single morning score makes obvious: recovery sliding for several days running, strain stacking well above baseline, resting heart rate creeping up. By the time it shows as burnout, a stalled plateau, or an injury, the member has often already lost momentum or blamed the app. The agent watches the rolling pattern against this member's own baseline and reaches out before the wall, naming what the trend shows and suggesting a deload or an easier stretch. It is preventive guidance, not diagnosis: it reads strain and recovery signals to protect training, and anything that looks like a health concern is handed to a clinical pathway.

How it works

  1. 1
    Trigger. a multi-day deviation from the member's baseline — recovery trending down across several days, sustained high strain, or a rising resting-heart-rate pattern synced from the wearable or health platform.
  2. 2
    Decision. the agent reads the trend (not one reading) against this member's personal baseline and recent training load, confirms it has not already raised it, and decides whether an early-warning nudge genuinely helps now or whether the pattern is still within their normal range; every unprompted message passes a quality and tone judge.
  3. 3
    Action. it sends one concise message naming the multi-day pattern and offering a concrete, non-clinical adjustment — a deload day, a lighter block, an extra rest day — framed as protecting progress rather than scolding, on the channel the member answers.
  4. 4
    Follow-through. the trend and the nudge write back to memory so the next message builds on it, the nudge is cancelled if recovery rebounds before it would land, and any reply that points to a medical concern (chest symptoms, injury, illness beyond fatigue) stops the guidance and routes to the clinical pathway.
Example
AgentHeads up — your recovery has trended down three mornings in a row and your strain's been running high the whole time. That's the pattern that usually comes before a wall, not after. I'd plan a deload: an easy session or a full rest day today, and keep this week lighter. Protecting the progress you've built, not undoing it.
Customermakes sense, I have been dragging
AgentThat tracks with the trend. Take it easy today and I'll watch whether recovery climbs back over the next couple of mornings — if it doesn't, we'll rethink the week together.

Configuration

How the agent is wired for this use case.

Triggera multi-day baseline-deviation flag (recovery_trend_down, sustained strain_high, rising-resting-hr pattern) computed over a rolling window from the wearable or health-platform feed.
Tools & actions
  • Wearable / device feed · ingest recovery, strain, and resting-heart-rate readings and detect a sustained deviation from baseline across days, not a single dip.
  • App backend / member memory · read the member's personal baseline, recent training load, and goal to judge whether the trend is genuinely meaningful for them.
  • Knowledge base · retrieve the program's approved, non-clinical deload and recovery guidance for the message.
  • Messaging channel · deliver one judge-gated early-warning nudge on the member's preferred surface and schedule a follow-up check on the trend.
  • Clinical-escalation pathway · hand off if a reply surfaces a medical symptom or a concern beyond training fatigue.
Autonomytrend detection, baseline comparison, and the single deload nudge run unattended under judge gating, quiet hours, and frequency caps. The agent gives training, strain, and recovery guidance only — no diagnosis, medication, or clinical instruction from a biometric trend; under a wellness-safety policy class, anything reading as a health concern is detect-and-escalate to a human, never handled by the agent.
Channelspush · chat · in-app
Escalationa reply indicating a medical symptom or injury, a reading pattern outside safe training-guidance bounds, or any request for clinical interpretation routes to the clinical pathway rather than being answered.

What you need

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

Signals

wearable recovery, strain, and resting-heart-rate readings with a rolling baseline, plus member replies.

Data

the member's personal baselines and recent training load, goal, channel and notification preferences, consent for health-data use, and persistent memory.

Guardrails

training and recovery guidance only — no clinical or medication advice from a biometric trend, with a stated boundary and mandatory clinical hand-off on any health concern; the warning fires on a sustained multi-day pattern, not one low reading; judge gating on every nudge; frequency caps and quiet hours so it supports rather than nags; stale-nudge cancellation when recovery rebounds; explicit consent and compliant handling of sensitive health data; visible opt-out.

Metrics it moves

  • churndown: catching strain-creep before burnout or injury keeps members training and subscribed through the stretch that would otherwise stall them.
  • csatup: an app that warns the member before they hit the wall feels like it is genuinely watching out for them, deepening reliance on the product.
  • dau-mauup: a timely, trend-tied early warning gives the member a reason to engage rather than ignore a passive dashboard.

Related use cases

Engage / Lifecycle

Wearable-driven rest and recovery nudge

the single-reading rest touch this extends into a multi-day trend-deviation warning

Prevent / Care

Pre-symptom illness alert

the sibling that reads biometrics for an illness-precursor pattern rather than training strain

Engage / Lifecycle

Proactive coaching companion for fitness and wellness

the broader habit-and-motivation coach this preventive warning lives alongside

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