Rapid Disenrollment Save

When someone drops a plan in the first few weeks, they usually didn't change their mind. Something broke. A card never came, a doctor wasn't covered, a drug cost more than they expected, a premium came out twice, or nobody explained something. This campaign catches those early drops the moment they happen and calls the member the same day.

The AI isn't there to talk anyone out of leaving. It's there to find out what happened, in the member's own words, and get them to your team, because whatever broke might be fixable and they might not know that. It has a twin, Win-Backs, that catches cancellations on policies older than 30 days. More on the split below. For the full playbook on this use case, see Save Cancelling Members.

In the messages below, anything in brackets fills in for each member: [First name], [AI name] (whatever you've named your AI employee), [Agency], [Agent name] (their assigned agent), [Agency phone], and [Carrier].

At a glance

How people get in

The campaign comes with an automation that watches for policy status changes. When a policy moves to Future Cancel, Terminated, or Cancelled, careCycle checks that policy's effective date:

So the two campaigns split your cancellations cleanly: a member can't land in both from the same status change. One thing to know: out of the box, MA Win-Backs only starts on Future Cancel, so an older policy that goes straight to Terminated or Cancelled isn't caught by either campaign unless you add those statuses to Win-Backs' automation. If the plan just started and it's already cancelling, that's this campaign. If they've been on it for a while, that's Win-Backs. If you want both groups covered, turn on both campaigns.

The date check looks at the policy that changed, not any other policy the member has. That last row matters for Future Cancel: if a member enrolled in a plan that hasn't started yet and it gets cancelled before it kicks in, the effective date is still in the future, so neither campaign picks it up. Same if the effective date is blank. You'd add those people yourself.

What the AI says

The campaign comes with three AI employees: one for daytime calls, one for after hours, and one for appointment callbacks.

Daytime calls (the main one)

The call opens with:

Hi, this is [AI name] at [Agency], I work with [Agent name]. Am I catching [First name] at an okay time?

Once they confirm, the AI gives the reason plainly:

Hey [First name], so uh I'm calling for [Agent name] because your health plan's showing as cancelled on our end, and it just started, so we wanted to check in on that.

Heads up: the AI says "showing as cancelled" for Future Cancel policies too.

Then it asks one open question and lets them talk:

Did something happen with it, or is this news to you?

That one question sorts the whole call. Whatever they say, the AI writes down the actual reason in their words (which drug, which doctor, which bill, which date), not just "member unhappy." Then it routes it:

• They didn't know (maybe a carrier error, or a cancellation they never asked for). It treats it as urgent without alarming them: "Ok yeah that's worth looking at right away. Can I get you over to someone on [Agent name]'s team who can pull it up?"
• Something went wrong and they cancelled because of it. It doesn't defend the plan. It says the problem is real and offers the team: "Yeah ok, that makes sense. Honestly that's the kind of thing [Agent name]'s team can usually sort out. Want me to get someone on real quick?"
• They moved to another plan on purpose. No fighting it and no bad-mouthing the other plan or agent. It notes it, makes one light offer of a review, and drops it if they say no.

If they won't take a transfer, it offers a callback, checks your calendar, and books a real day and time.

It's also ready for the stuff people say on these calls:

• "Why is it cancelled?" It's honest that it doesn't know, all it can see is that it's showing cancelled, and offers the team, who can see the actual reason.
• "Am I covered right now or not?" It won't guess either way, because being wrong could stop someone from getting care. It says the team can confirm it properly and treats this as the most urgent version of the call.
• They're angry. It lets them finish, doesn't defend anybody, and gets the complaint down in their own words.
• "I never signed up for this." It takes it seriously, notes it exactly, and routes it. No arguing.
• A premium came out twice. It doesn't try to explain billing. It notes the details and routes it.
• "Are you a robot?" "I'm a virtual assistant, I work with [Agent name]."
• "Is this a scam?" It mentions their agent, your agency, and when their coverage started, and gives them your number to call and check.
• "Take me off your list." It confirms they'll be removed and lets them go. No save attempt.

It never gives plan advice, never quotes a cost, and never says whether they can re-enroll. That's your team's job.

After hours

Same opener, same question, same notes. The difference is nobody's around to take a live transfer, so instead of transferring, the AI makes sure someone on the agent's team follows up, and confirms the best way and time to reach them. If someone calls in after hours, it lets them know everyone's away from their desk, finds out what they're calling about, and takes down the details.

Appointment callbacks

When an appointment comes up, this employee makes the callback. It mentions the recorded line, says it's calling for their agent about the scheduled appointment on their [Carrier] plan, checks that now still works, and transfers them to the team. If now's bad, it finds a new time on your calendar and books it.

The cadence

Everything happens on Day 1, about six hours apart. Early drops are time sensitive, so this one doesn't wait around.

Day 1: Call

Text, a few minutes before the call:

Hey [First name], it's [Agency]. Something's showing on our end with your [Carrier] coverage that we'd like to go over with you. Giving you a ring in just a minute.

Voicemail if they don't pick up:

Hi [First name], this is [AI name] calling from [Agent name]'s team at [Agency]. Something's showing on our end with your [Carrier] coverage that we'd like to go over with you, and it's the kind of thing that's better sorted sooner than later. Nothing you need to worry about. Give us a ring back at [Agency phone] when you have a minute and we'll look into it together.

Day 1: Text (if we didn't reach them)

Sorry we missed you just now. We're seeing something on your [Carrier] coverage and want to make sure nothing slipped through. We'll try you again, or reach us at [Agency phone] anytime.

Day 1, about six hours later: Call

Text, a few minutes before the call:

[First name], trying you again in a couple minutes about what we're seeing on your [Carrier] coverage. The sooner we look at it, the better.

No voicemail.

Day 1: Text (if we didn't reach them)

Missed you again. Nothing to worry about, we'd just like a few minutes with you about your [Carrier] plan. You can reach us at [Agency phone] if that's easier.

Day 1, about six hours after that: Call

Text, a few minutes before the call:

Still hoping to catch you. One more try coming your way in a minute so we can take a look at your [Carrier] coverage together.

Voicemail if they don't pick up: the same one as the first call.

Day 1: Text (if we didn't reach them)

Looks like we keep missing each other. No worries at all. If anything's off with your [Carrier] coverage, we're at [Agency phone] and glad to help sort it out.

After that, the journey ends.

Calls and texts only go out during your business hours and between 8 AM and 9 PM in the member's time zone. So if a later call lands outside that window, it waits until you're open again.

How each call works

• If they get transferred, book an appointment, or the AI logs an inquiry, they're done. The journey ends right there. Booking an appointment ends all of that member's active campaigns.
• If they talk to the AI (including telling it they want to cancel, that they're confused about the policy, or that their package never showed up), they're done too. No more calls.
• If we can't reach them (no answer, voicemail, "call me later" without booking, or the call drops right away), they get the follow-up text and move to the next call.
• If it's a bad number, or they ask not to be called, they're stopped right away.

Before you turn it on

• Set up your escalations in Employees > Escalations. Point Transfer Call at a transfer destination and Book Appointment at an appointment configuration. The campaign can't be activated until both point somewhere. Inquiries go to your Inbox and the Inquiries page automatically, and that's where the member's reason for leaving shows up.
• Check your member data. The AI needs each member's first name, last name, carrier, and assigned agent, plus your agency name. People missing any of these get skipped, and so does anyone without a phone number or a time zone or state. It also uses the policy's effective date when it's there.
• Make sure your policy statuses stay current. This campaign only fires when a status changes in careCycle, so your policy data has to keep flowing in.
• Check your caller ID. Out of the box, this campaign calls from your agency's main number (the one your Receptionist answers), so make sure you have one. You can see this on the Phone Numbers tab: Use organization main number for outbound calls is switched on. Want it to call from its own number instead? Switch that off and add a number there (you can only change it while the campaign isn't active). See Managing Your Phone Numbers.
• Get SMS registration done so the texts can go out. See SMS and RCS Registration.
• Test it. See Test Your Campaign.
• Click Activate on the Overview tab. Then add anyone who's already in a cancelling status, since the automation won't go back for them.

Changing any of this

• What the AI says on calls: the Employees tab. See Edit Your AI with Employees.
• Texts, voicemails, and timing: the Pulse tab. Click any call or text step to edit it. See Pulse.
• Who gets in: the Automations tab. That's where the statuses and the 30-day effective date rule live. See Automations Explained.