Florida Insurance Producer Leads — Newly Licensed & Unappointed Agents

Recruiting-ready contact data for Florida insurance producers, built from the Florida Department of Financial Services Licensee Search bulk downloads — the state's own public-record files.

Two things this Actor does that a licence-lookup tool cannot: it finds the producers who just got licensed, and the ones who are licensed but no carrier has appointed yet. Those are the two lists agent-recruiting teams actually buy.

What does this Actor do?

• Downloads the official All Valid Licenses – Individual file (321 MB, 1,230,054 licence rows, 968,206 distinct producers — measured on the 2026‑07‑27 file) and filters it for you.
• Newly licensed — producers whose first-ever Florida licence falls in your date window. Not "someone who added a licence class last month": the Actor reads every licence each person holds before deciding. Measured on the live file: 8,930 new entrants in the last 30 days, 112,901 in the last 12 months.
• Licensed but not appointed — joins the state's three active-appointment files (3,127,292 active appointments) by licence number and keeps only producers with none. Measured: of 30,954 producers licensed in the last 90 days, 12,070 (39%) have no active carrier appointment.
• Search all licences — full control: licence class, county, city, state, residency, contact requirements.

Why use it?

	This Actor	Typical FL licence lookup Actor
Filter by licence issue date	✅	❌
Find producers with no appointment	✅	❌ (they verify appointments instead)
One row per person, licence classes nested	✅	❌ raw rows — the same person 1.27×
Miami-Dade spelling accepted	✅ (register spells it Dade)	❌ returns 0 rows
Masked empties (N/A, 0000000000) → null	✅	❌ counted as filled

How to use it

1. Pick a Lead preset — Newly licensed producers is the default.
2. Set the date window (defaults to the last 90 days for that preset).
3. Optionally narrow by licence category, county/city, residency, or require an e-mail/phone.
4. Run. A run reads the whole state file, so expect ~3 minutes regardless of how many rows you ask for (~12 minutes when the appointment join is on).

Input example

{
  "preset": "newly_licensed",
  "datePreset": "last_60_days",
  "residentOnly": true,
  "licenseCategories": ["life_and_health", "health"],
  "requireEmail": true,
  "maxResults": 2000
}

Unappointed-producer recruiting list:

{
  "preset": "no_active_appointment",
  "datePreset": "last_90_days",
  "requireEmail": true,
  "maxResults": 5000
}

Output example

{
  "licenseNumber": "A276085",
  "npn": "636278",
  "fullName": "Keith J Walker",
  "firstName": "Keith",
  "middleName": "J",
  "lastName": "Walker",
  "email": "inssol1@earthlink.net",
  "businessPhone": "2394037420",
  "residencyType": "Resident",
  "isFloridaResident": true,
  "businessAddress1": "INSURANCE SOLUTIONS",
  "businessAddress2": "2171 PINE RIDGE RD SUITE D",
  "businessCity": "NAPLES",
  "businessState": "FL",
  "businessZip": "34109",
  "businessCounty": "Collier",
  "mailingAddress1": "INSURANCE SOLUTIONS",
  "mailingCity": "NAPLES",
  "mailingState": "FL",
  "mailingZip": "34109",
  "firstLicenseDate": "1993-06-12",
  "latestLicenseDate": "1995-05-20",
  "licenseClassCount": 2,
  "dataAsOf": "2026-07-27",
  "licenseTypes": ["GENERAL LINES (PROP & CAS)", "LIFE INCL VAR ANNUITY & HEALTH"],
  "licenseCategories": ["property_casualty", "life_and_health"],
  "licenses": [
    { "licenseTypeCode": "0215", "licenseType": "LIFE INCL VAR ANNUITY & HEALTH", "licenseCategory": "life_and_health", "licenseStatus": "VALID", "licenseIssueDate": "1993-06-12" },
    { "licenseTypeCode": "0220", "licenseType": "GENERAL LINES (PROP & CAS)", "licenseCategory": "property_casualty", "licenseStatus": "VALID", "licenseIssueDate": "1995-05-20" }
  ]
}

With appointmentFilter on, every row also carries hasActiveAppointment, activeAppointmentCount and appointingEntities (up to 10 carrier names).

Data fields & measured coverage

Coverage measured two ways: a complete parse of a full 1,189,056-row file (population), and a live 5,000-producer platform run (sample). Both are shown — no estimates.

Field	Full-file	Live sample	Note
email	99.99%	100.00%	lower-cased, regex-validated; invalid → null
businessPhone	99.93%	99.60%	10 digits; 0000000000 placeholders → null
npn	99.86%	99.98%	National Producer Number — your CRM/AMS join key
fullName / firstName / lastName	100%	100%	rebuilt in natural order for mail-merge
middleName	61.4%	84.8%
licenseTypes / licenses[]	100%	100%	78 Florida licence classes, code + description
firstLicenseDate / latestLicenseDate	99.99%	100%	ISO dates across all classes the person holds
licenseClassCount	100%	100%	22% of producers hold more than one class
businessCity / State / Zip	~99.9%	~99.9%
businessCounty	96.2% of Florida-address rows (39.6% of all rows)	55.0%	the state records a county only for Florida business addresses
businessAddress2	22.8%	40.1%	suite/unit line; duplicates of line 1 are dropped
mailing address block	~100%	~100%

Honest notes, because they change what you can ask for:

• businessCounty is empty for non-residents. Within Florida business addresses it is 96.2% filled (450,910 of 468,831 rows measured), but 59% of Florida licences are held by out-of-state producers and the state leaves their county blank — so county filters only make sense for Florida addresses.
• Licence status is effectively constant. The source file is "all valid licences": 99.98% are VALID, 0.02% VALID - PROBATION. It is exposed as a filter and inside licenses[], not as a headline column, because it tells you almost nothing.
• No profile URL is returned. The state's licensee search is a form-driven page with no stable per-producer URL that could be verified row by row, so raw identifiers (licenseNumber, npn) are shipped instead of a link that might not resolve.
• "Newly licensed" is measured, not inferred, but a small number of re-issued licences move their issue date forward: comparing two real snapshots nine months apart, 99.94% of issue dates were unchanged and 0.06% had been re-issued.

Freshness

The state rebuilds these files daily. Measured: the live file carried Last-Modified: Mon, 27 Jul 2026 06:21:02 GMT on a 2026‑07‑28 run, and two independent historical snapshots each contained licences issued 1–2 days before the file was published. Every row carries dataAsOf so you can prove the vintage of a list you exported.

Pricing & cost

A run reads the entire state file, so cost is dominated by that fixed scan, not by row count. Measured on the platform at 4 GB:

Run	Rows	Time	Compute units
Smoke	25	183 s	0.203
Filtered (Dade, life & health)	2,000	187 s	0.208
Adjusters, e-mail + phone required	6,000	192 s	0.213
Newly licensed, 12 months	10,000	198 s	0.220
With appointment join (~1.3 GB)	2,000	748 s	0.831

Peak memory stayed at 250–430 MB; 4 GB is the default only so that very large unfiltered pulls do not run out of room. Pull bigger batches per run — 10,000 rows cost barely more than 25.

Tips

• Recruiting cadence: run newly_licensed + last_30_days weekly; you will see roughly 8,000–9,000 new producers a month statewide.
• The sharpest recruiting list is no_active_appointment + last_90_days + residentOnly — newly licensed, in-state, and nobody has signed them yet.
• Use Miami-Dade freely; it is translated to the register's Dade.
• Set maxResults: 0 for the whole register, and raise the run's memory to 8 GB for that.
• Turning One row per producer off gives you the raw licence-level rows if your CRM prefers them.

FAQ & legal

Where does the data come from? The Florida Department of Financial Services publishes these files at licenseesearch.fldfs.com/BulkDownload (the CSVs themselves are served from www.myfloridacfo.com). Florida licensee contact information is public record under Florida law.

Does it need an API key or proxy? No. No credentials, no proxy, no browser.

Is this personal data? Yes — these are real people's business contact details, published by the state. You are responsible for how you use them. Outreach to these contacts is subject to CAN-SPAM, TCPA and Florida law; that obligation sits with you as the sender, not with this Actor. Do not use the data for anything the source's public-record purpose does not support.

What if the state changes the files? The Actor validates the header of every file it downloads and aborts with a clear message if a column disappears, if it receives HTML instead of CSV, or if a download is truncated. It never emits rows that look complete when a layer failed.

Why does a 25-row run take three minutes? Because the answer to "who is newly licensed" requires reading every licence in the state. There is no server-side query to push the filter into.

Changelog

0.1 (2026-07-28)

First release — see CHANGELOG.md.