All use cases

Onboard / Activate

Account-recovery to preserve cycle history

She is locked out and one tap from starting over — the agent recovers her old account so the years of cycle history that power her predictions don't vanish.

Next best thingClose the loopBehavioral triggerRisk triggerChatIn-appWellness (B2C apps)Subscription

What it does

When a femtech user can't log back in, the easy path is to create a fresh account, and that quietly destroys the thing the product is built on. Cycle predictions, phase insights, and trend reads all depend on months and years of logged history, and a duplicate account starts that clock from scratch — so the user lands back in a worse, less accurate app and the provider carries an orphaned record and a churn risk it can't see. Instead of letting a stuck login become a silent restart, the agent recognizes the recovery moment, verifies identity securely inside the chat, and walks her back into her original account with her history intact.

How it works

  1. 1
    Trigger. a login-trouble signal — a failed sign-in, a "can't access my account" message in chat, or the user beginning a fresh sign-up on a device that already matches an existing record.
  2. 2
    Decision. the agent checks for a likely existing account behind the attempt, recognizes the duplicate-creation risk, and routes into secure identity verification rather than waving her toward a new signup.
  3. 3
    Action. it runs a secure in-chat verification step (a one-time code or token-based check handled by the app backend, never credentials typed to the agent), confirms the match, and restores access to the original account with its full logged history.
  4. 4
    Follow-through. the recovery and its outcome write back so the duplicate is prevented before it is created; an unverifiable or higher-risk case is handed to a human support agent with context rather than force-resolved, and a sensitive-data or security concern routes to the right team.
Example
CustomerI can't log in and I'm about to just make a new account — I'll lose all my tracking, won't I?
AgentYou would, and there's no need. It looks like you already have an account with us — let's get you back into that one so your history stays. I'll send a one-time code to your email to confirm it's you; nothing to re-enter here.
Customerok got the code, it's 4-1-9-2-0-3
AgentThat's you. You're back in your original account with all your logged cycles intact — your predictions pick up exactly where they left off.

Configuration

How the agent is wired for this use case.

Triggera login-trouble or duplicate-signup-risk signal — a failed sign-in, an account-access request in chat, or a fresh sign-up starting against a device that matches an existing record.
Tools & actions
  • App backend · check for an existing account behind the attempt, run the secure verification, and restore access to the original record
  • Identity/auth system · issue and validate a one-time code or token-based verification; credentials are never typed to the agent
  • Memory store · recognize the recovery context and write the outcome back so the duplicate is prevented, not just merged later
  • Messaging channel · run the recovery conversation in chat or in-app and confirm restored access
  • Knowledge base · ground the recovery steps and the "why your history matters" explanation
Autonomyidentity verification and account restoration run only on a successful secure check handled by the auth system; the agent never sees or stores raw credentials, never overrides a failed verification, and a higher-risk or unverifiable attempt is escalated rather than resolved autonomously.
Channelschat · in-app
Escalationan attempt that fails verification, looks like an account-takeover risk, or raises a sensitive-data concern hands off to a human support agent with full context — the agent surfaces the case, a person makes the access call.

What you need

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

Signals

failed-sign-in and account-access-request events, plus a duplicate-signup-risk signal when a new registration matches an existing record

Data

existing-account lookup keyed to a verifiable identifier, verification state from the auth system, consent state, and memory of the recovery attempt

Guardrails

secure identity verification before any restore; raw credentials never handled by the agent; escalate on failed verification or takeover risk; sensitive-data handling and regional privacy limits; a human owns the final access decision on ambiguous cases

Metrics it moves

  • activation-rateup, because a locked-out user is recovered into a working account instead of restarting cold in a degraded one
  • churndown, as the longitudinal history that makes predictions accurate (and the app worth keeping) is preserved through the lockout
  • csatup, since "I almost lost everything" becomes a one-conversation recovery instead of a support ticket and a fresh start

Related use cases

Collect / Recover

Self-serve subscription cancel and refund in chat

the same secure in-chat identity gate applied to a billing action

Collect / Recover

Data-request self-service for sensitive PII

the sibling self-service path handling sensitive-data requests behind verification

Onboard / Activate

In-app activation companion

the general activation surface this protects from a duplicate-account restart

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