Made with love and nicotine in San Francisco
Insights
September 21, 2026
Vladimir Taikov

AEP 2027: the season guide for Medicare agencies

An older couple walking a wooded park path, one using a rollator. Overlaid text reads 2027 AEP Season Plan, with the careCycle logo.

AEP is 54 days. Your agent count is fixed. Your book isn't getting smaller.

Most agencies solve that by picking one: protect the book, or write new business. Then they spend Q1 backfilling whatever leaked.

Here's what changed for plan year 2027, and the order that keeps you from having to choose.

The short version

  • One pass at your book can eat a third to two-thirds of the season. Before you write a single new app.
  • Carriers grade you on accretion above 80% and rapid disenrollment under 10%. Those numbers decide your contracts, not just your commissions.
  • The apps you write in October carry the most risk. Eleven weeks to effectuation, seven of them inside AEP, where anybody can rewrite them.
  • Work the book you own before you buy leads. The order is the strategy.
  • Your book stays in motion until March 31. Most agencies stop working it on December 8.

The capacity problem nobody puts a number on

Every AEP plan is a capacity plan. Most never get written down as one.

So run it on a 20-agent shop with 5,000 members.

What you've got:

  • 20 licensed agents
  • About 46 working days, at six-day weeks
  • About six real selling hours a day, after huddles, breaks and wrap
  • Roughly 5,500 agent hours. That's the whole season.

What one pass at the book costs:

  • 5,000 members, 30 minutes each including wrap
  • 2,500 hours. Nearly half of everything you have.

Change the inputs and the answer moves a lot. Thirty agents against 3,000 members lands near a fifth. Fifteen agents against 8,000 members runs out of season before it finishes the book. For most agencies it sits somewhere between a third and two-thirds.

And that's if you reach everyone on the first dial. You won't. At a 25% contact rate, 5,000 conversations takes about 20,000 attempts.

The dialing isn't the expensive part. The expensive part is paying licensed, certified, carrier-appointed producers to listen to ringing.

Then there's the volume you never staffed for. Across four million Medicare member service calls, 26%+ of qualified inbound opportunities came in outside business hours. You paid to generate those. In October they hit voicemail.

That's why agencies pick growth over retention in October. Not strategy. Arithmetic.

What changed for plan year 2027

Six changes came with the CY2027 final rule. The marketing provisions kicked in October 1, 2026, so they were live before this season opened.

  • The 48-hour SOA wait is gone. Capture scope and run the appointment on the same call. The SOA itself didn't go anywhere.
  • The TPMO disclaimer moved, it didn't disappear. It no longer has to hit in the first 60 seconds. It does have to hit before you talk benefits.
  • SHIP came out of the disclaimer. Medicare.gov and 1-800-MEDICARE stay in.
  • Sales and marketing recordings: six years, not ten. Full audio the first three, audio or transcript for four through six.
  • Back-to-back events are allowed. The 12-hour gap at the same venue is gone, as long as you tell attendees when marketing starts and let them leave. You can collect SOAs at educational events now too.
  • Superlatives are back, with documentation. The ban on misleading claims didn't move.

Read that fourth one twice. Six years covers sales and marketing calls. Anything that produced an enrollment still carries ten. Separate those buckets before you touch retention settings.

The SOA change is the one that reorganizes your season. Renewal outreach that used to book an appointment 48 hours out can warm-transfer on contact. Every appointment you don't have to set is an appointment nobody no-shows.

Sources: CMS issued the rule (CMS-4208-F3 and CMS-4212-F) April 2, 2026 and published it April 6. See the CMS fact sheet, the rule itself at 91 FR 17384, or our change-by-change breakdown.

The two numbers your overrides depend on

Volume stopped being the whole scorecard a while ago. Carriers grade what happens after the app now. Leading FMOs tell us the targets sit around two numbers.

Accretion, above 80%. How many of your apps actually effectuate. One that never becomes a paying member still cost you an agent hour.

Rapid disenrollment, under 10%. How many members leave right after they join. Definitions vary by carrier, so check what your hierarchy actually measures.

Miss these and it isn't a commission problem. It's a contract problem. Your overrides, your bonus tiers, your access to the better contracts: all of it sits on standing.

Why October applications behave differently

Write a member on October 16 and they don't have coverage on October 17. They've got a plan that starts January 1. Eleven weeks out.

Seven of those weeks are inside AEP. Seven weeks where anybody can rewrite them, and everybody working the same disruption list is trying to.

Seven weeks is also when your agents move fastest and confirm the least. Which is exactly what you're paying them to do in October.

So look at those two metrics again. The apps most likely to drag both down are the ones you wrote first, in the weeks your board looked best.

It isn't complicated. Longer gap to effectuation, more exposure to re-shopping, less confirmation at the point of sale.

None of this says write less in October. It says build the post-enrollment layer before October, so volume and accretion stop trading against each other.

The order of operations

Six phases. The sequence is the strategy, and running them out of order is what buys you a Q1 of backfilling.

  1. Before October 15 · Get the book in one place. Every policy with a carrier, an effective date and a writing agent NPN. Reassign orphans. Segment by exits, disruptions and benefit cuts. You can market and book from October 1, so those two weeks are for filling October 15, not warming up.
  2. October 15 on · Existing clients first. One pass per member. Renewal, plan change and ancillary fit in one conversation instead of three.
  3. Then acquisition. Not before. A net-new lead won while an existing member goes uncalled is a trade you lost twice. Once on the churn, once on the media you bought to replace them. And pre-screen third-party traffic before a licensed agent picks up, because publishers get paid for volume, not intent.
  4. Post-enrollment, while your agents keep selling. The welcome call is a retention asset, not a courtesy.
  5. January 1 to March 31. Members can still switch. Most guides stop at December 7.
  6. After the dust settles · Attachment rate. A member holding one product is easy to move. A member holding four is anchored.

Phase 5 is where books quietly leak. A member hits the pharmacy, finds a drug isn't covered the way they expected, pays more than they planned. They don't call you. They decide you got it wrong, and in February they take somebody else's call.

That's the sequence. The guide is how you run it.

Nineteen pages with everything this post leaves out:

  • Segmentation criteria for every bucket
  • Warm-transfer routing rules
  • The full six-tier source hierarchy for October media buying
  • Post-enrollment cadences by lead source and plan type
  • A 28-point readiness checklist, month by month

Get the AEP 2027 guide → Free, and we'll email it straight to you.

What it looks like when it runs

  • Double-digit drops in CPA, from pre-screening that stops paying producers to talk to people who were never shopping.
  • Double-digit lifts in effectuation, which is the accretion number your overrides ride on.
  • Over $2M in AEP revenue driven for customers last season.

Results vary by agency. What doesn't vary much: the unit economics improve once the post-enrollment system is running, because you stop paying twice for the same member.

Common questions about AEP 2027

How many agents do I need for AEP? Work backward from capacity, not headcount. Book size times handle time gives you the hours for one pass. Put that against agents times selling hours times working days. The guide has the worksheet.

What happens to my book when a producer quits mid-season? Their members go orphaned, and orphans are the highest-churn segment you own, because nobody's accountable for them. Reassign by NPN before October 15, not in January when the disenrollments show up.

Won't automated outreach to seniors feel like spam? Spam is untargeted. A call that opens with a member's first name, their plan and their writing agent's name isn't spam, and they can tell inside a sentence. The segmentation is the whole difference, which is why it gets its own section in the guide.

Is it worth working the book after December 7? That's where the cheapest saves are. Anyone already in an MA plan gets one more change through March 31, and the members who leave in Q1 mostly decided in January that you got it wrong. Nobody calls to tell you.

Do I still need an SOA? Yes. CMS killed the 48-hour wait, not the requirement.

How long do I keep call recordings? Six years for sales and marketing, full audio the first three. Enrollment calls still carry ten.

Where to start

Already a client? Talk to your account lead about segmenting your book and building campaigns for the season.

Not yet? Start with the guide. It's the operating document behind everything above, and it's free.

Get the AEP 2027 guide →

For agencies with more members than hours.

Stop losing members
to competitors today.