All use cases

Engage / Lifecycle

Financial-goal commitment check-in

When a customer tells the agent "I'll move $200 to savings this month," the agent remembers — and comes back to ask how it went.

Close the loopTime-based triggerBehavioral triggerPushChatIn-appEmailFintech

What it does

Money goals are set in a moment of intent and then quietly abandoned: a customer says they'll start a savings habit, cap their dining spend, or chip away at a card balance, and the app never mentions it again. The promise evaporates, the behavior never forms, and the banking relationship stays shallow and easy to leave. The agent writes the commitment into the customer's memory the moment it is made, returns at the right time to ask how it is going, and celebrates the win or helps re-plan the miss, so a stated intention becomes a habit that deepens the primary-account relationship.

How it works

  1. 1
    Trigger. a conversation ends with a financial commitment ("I'll set aside $200 this month," "I'll keep dining under $300"), or a goal-related thread is dropped before it resolves.
  2. 2
    Decision. the agent extracts the commitment, writes it to the customer's memory, and schedules a check-in for the moment it can be judged — about 24 hours for a one-off action, the period's end for a budget. It reads account signals to see whether the goal was actually met before it writes a word. Every message passes a quality judge before it sends.
  3. 3
    Action. a short, specific message on the customer's preferred channel that picks up the exact goal — "How's the savings goal going? You're $140 of $200 there" — and opens straight back into the conversation, celebrating progress or offering to adjust the plan.
  4. 4
    Follow-through. if the customer hits the goal or returns on their own first, the pending check-in is cancelled so nothing arrives stale or wrong; the outcome writes back to memory so the next conversation builds on it, and impact is measured against a holdout.
Example
AgentA few weeks ago you set yourself a goal to move $200 into savings this month. You're at $140 with a week to go — nicely on track. Want me to set up the last transfer so you can confirm it and finish the goal?
CustomerYeah let's do it, I keep meaning to and forgetting.
AgentI've teed up the transfer in your app — just authorise it on the secure screen and the goal's done. I'll set the same target for next month and check in again.

Configuration

How the agent is wired for this use case.

Triggera commitment_captured event from a conversation (a stated savings, budget, or debt-paydown goal), or a goal_thread_dropped event when a money-goal conversation ends unresolved.
Tools & actions
  • Customer/account model · write the commitment, its target, and its deadline; recall it at check-in time
  • Core banking platform · read balances, transfers, and category spend to verify whether the goal was actually met before messaging
  • Scheduling/timer service · set the check-in for the right horizon (next-day action, period-end budget) and cancel it on early goal completion or customer return
  • Money-movement service · set up the transfer that finishes the goal and hand off to the bank's own secure, authenticated step for the customer to authorise; read the confirmation back, never move funds itself
  • Messaging channel · deliver the personalized check-in on the customer's preferred surface
Autonomycapturing commitments, scheduling, and sending the judge-gated check-in run unattended under quiet hours and frequency caps. Any actual money movement (completing a transfer, adjusting a standing order) is never autonomous and the agent never moves funds itself: under a money-movement policy class it orchestrates up to the bank's own secure, authenticated step, which the customer authorises in-app, and reads the confirmation back.
Channelspush · chat · in-app · email
Escalationa stalled goal paired with hardship or distress language, a dispute, or any request beyond a self-serve transfer hands off to a human with the goal context attached.

What you need

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

Signals

conversation-end events carrying the extracted financial commitment, dropped-thread events, account events (transfer completed, spend against a budget), and customer-return events to cancel pending check-ins

Data

the customer model with goal entries and deadlines, the account history needed to judge progress, channel preference and consent, quiet hours, persistent memory of prior goals and outcomes

Guardrails

judge gating on every check-in before it sends; money movement only on the bank's own secure, authenticated step that the customer authorises, never executed by the agent, under a money-movement policy class; verify goal status from account data before messaging so no false "well done" fires; cancellation on early completion or customer return; consent, quiet hours, and frequency caps; holdout measurement; the client keeps owning the send channels

Metrics it moves

  • reactivation-rateup: a personal, well-timed check-in on a goal the customer set themselves brings them back into the app where a generic balance notification is ignored
  • dau-mauup: goal check-ins open into real conversations and money actions rather than dead-end pushes
  • ltvup: customers who form a saving or budgeting habit treat the account as their primary one and stay longer, measured against the holdout

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