On this page
In most sales, timing is a tactic. In Medicare, timing is the product. The whole business runs on a small set of federal dates, and a beneficiary who reaches a broker in the wrong week can be locked into the wrong plan until the next enrollment window comes around.
This piece draws on a system we built and ran for a Medicare agency. That engagement has since ended and the agency moved to a different CRM, so everything here is written in the past tense on purpose. The lessons about how this vertical works did not leave when the engagement did.
Related case study
Paper intake became a searchable, trackable Medicare lead operation
Capstone Health is an independent Medicare agency in Tampa. Every inbound call started life on a paper lead sheet, and everything after it depended on somebody remembering. We rebuilt intake, timing, consent, and follow-up into one operating system, and audited 61 workflows and 116 tags on the way through.
See our workThe four windows that run a Medicare year
Four federal windows set the rhythm of the whole book. T65 is the month a person turns 65 and first becomes eligible. AEP runs October 15 to December 7. The scope-of-appointment agreement has to be in place 48 hours before a sales meeting. The annual notice of change lands September 30.
Four federal windows set the rhythm of the entire book. Intake either knows them cold or spends the year fighting them.
| Window | When it is | Why intake has to know it cold |
|---|---|---|
| T65 | The month a person turns 65 | First eligibility, and the cleanest signal in the book. You can see it coming months out, per lead, straight from a date of birth. |
| AEP | October 15 to December 7 | Annual enrollment, when current members can change plans. The busiest stretch of the year, and it arrives on a date visible a full year ahead. |
| SOA | 48 hours before a sales meeting | The scope-of-appointment agreement on what may be discussed. It has to be in place before the appointment, never captured after it. |
| ANOC | September 30 | The annual notice of change tells members what is shifting on their plan, and it starts the clock on the questions that follow. |
None of this is abstract. A lead who reaches the office two weeks into December has narrower options than the same lead in October, and intake either tracks that difference per person or quietly loses money on it.
Why paper intake survives, and where it fails
Paper is genuinely fast while you are on the phone, and capture is the part it does well. Everything after is where it breaks: re-entering the lead, classifying it against the right enrollment window, and remembering who needs a call back and when. In a business where timing is the product, follow-up that runs on memory is the whole exposure.
Every inbound call at the agency started on a paper lead sheet, and here is the honest part: that works. Paper is fast when you are on the phone, and a lot of good brokers live on it. The capture is the part paper does well. Everything after it is where paper breaks: re-entering the lead, classifying it against the right enrollment window, and remembering who needs a call back and when. Lead status lived across paper, memory, and scattered notes.
In a business where timing is the product, follow-up that depends on memory is the whole exposure. And unlike most industries, the stress test is scheduled: AEP volume arrives on a date you can see on a calendar a year in advance. The system we built moved every lead from paper to a structured record the same day, classified it against its real enrollment timeline, T65, AEP, open enrollment, dual-eligible status, and started the correct follow-up path immediately. Paper could still exist. It stopped being the only place the lead lived.
The flag that outranks everything: do-not-contact
One rule came out of this build and now runs on every CRM we touch: search first, never blind-write. You find the existing person before you change a single field.
Purchased lead lists are a fact of life in this vertical, and they carry a danger nobody talks about. On the book we rebuilt, 4,742 purchased contacts carried a do-not-contact flag. Deduplication, merging one person's multiple records into a single contact, is normally a hygiene chore. Here, dropping even one of those flags in a careless merge crosses from hygiene into compliance, and that is a different order of mistake.
So the priority was set deliberately: protecting that flag mattered more than resolving every last duplicate. The system proposed merges and a person approved them, and designing with that priority is the line between automation that helps a regulated business and automation that quietly builds its next violation.
Kill switches, and the fear that turned out to be correct
When a broker tells you what they are afraid of, that is not resistance to technology. That is the requirements document.
The owner was genuinely afraid that dispositioning a test lead would fire a real thanks-for-enrolling email to a live client, so he seeded his test records with fake email addresses. He was right to be careful. He had also flagged the risk of pushing Social Security and Medicare numbers through generic automation before we raised it ourselves.
Both fears turned into hard constraints. A lead disposition could never trigger an outbound message. The instant an appointment booked, every sequence attached to that lead stopped, so nobody got chased for something they had already done. And sensitive identifiers stayed out of the automation path completely, no Social Security number and no Medicare number, at any point.
What must stay human, permanently
Every plan recommendation and every enrollment belonged to the licensed agent, in the agent's own enrollment system. The automation never recommended a plan, never enrolled anyone, and never ran a benefit conversation. In a regulated vertical that separation is the only version that survives an audit.
Every plan recommendation and every enrollment belonged to the licensed agent, in the agent's own enrollment system. The automation never recommended a plan, never enrolled anyone, and never ran a benefit conversation. That separation was the design from day one, and in a regulated vertical it is the only version that survives an audit.
He trusted his own close rate, and he was right to. The hard part sat upstream of the close: getting a qualified person onto his calendar on the correct day of a window that shuts. He wanted the record pushed to his enrollment system the instant it booked, with the real outcome recorded after, and he was right about that too.
The checklist for any broker considering this
Five questions tell you fast whether a vendor has worked this vertical. Does the system know your compliance calendar? Can dispositioning a lead ever fire a message? What happens to a do-not-contact flag during a merge? Where do Social Security and Medicare numbers travel? When an appointment books, does everything that should stop actually stop?
If you run a Medicare book and you are weighing intake automation, forget the AI feature list. Five questions actually matter, and how a vendor answers them tells you fast whether they have worked in this vertical.
- Does the system know your compliance calendar, or does it treat a T65 lead like a generic funnel stage?
- Can dispositioning a lead ever fire an automated message, under any condition?
- What actually happens to a do-not-contact flag during a merge?
- Where do Social Security and Medicare numbers travel, and where are they kept out?
- When an appointment books, does everything that should stop actually stop?
A vendor who cannot answer those five in specifics has not run this kind of book. The insurance industry page covers how we think about the whole operating loop, and the full case study shows what this looked like in production.
The honest note about this engagement
The system ran live for months: 61 workflows and 116 tags audited and rebuilt, every inbound lead captured and routed digitally the same day. The agency has since moved to a different CRM and the engagement ended with it. We say so because the alternative is an article that quietly implies a client we do not have.
The system ran live for months: 61 workflows and 116 tags audited and rebuilt, every inbound lead captured and routed digitally the same day, and an automated nurture engine running from first contact through enrollment season. The agency has since moved to a different CRM, and the engagement ended with it.
We are telling you that because the alternative is an article that quietly implies a client we do not have. What the work produced is still real, and the compliance vocabulary in this piece came out of that build, not out of a content calendar.
Next step
Running a Medicare book on paper and memory?
Start with one question: does your intake actually track your enrollment calendar, or just hope to remember it? On a discovery call we walk your real windows together and mark which state changes are safe to automate and which stay with the licensed agent. The compliance lines above are where we begin, not where we negotiate.
Written by
Christopher J. Moreno
Chris is a solo AI consultant with five documented systems across construction, roofing, and Medicare insurance, every number on them measured before it was published. He builds operating systems for real businesses that need cleaner intake, clearer follow-up, and less invisible admin drag.
Our methodology
The Flo OS in practice
The approach behind this work follows the four phases of Flo OS, our operating methodology for turning messy business workflows into systems that run cleanly and compound over time.
See how we work →Related reading
Keep reading.
Explainer · 11 min
What an AI-Legible Business Actually Looks Like
Most businesses do not have an AI problem first. They have a workflow legibility problem. Here is the full diagnosis, the five-point checklist to score yourself, and the 90-day path out.
Proof · 7 min
How a Medicare Insurance Workflow Goes From Paper to Searchable, Trackable, and Automatable
A practical look at how an insurance workflow becomes easier to manage once intake, routing, and appointment logic live in a system instead of on paper and memory.
Operations · 7 min
How to Reduce Admin Overhead Without Adding Staff
Admin overhead in most small businesses is not visible as a line item. It shows up as hours that disappear into coordination, re-entry, and manual follow-through.