The List: What a Jail Phone Vendor Actually Keeps

Six categories, one account, nineteen months — the complete records demand filed against a jail phone vendor.

2026-11-13 · The List: What a Jail Phone Vendor Actually Keeps · Criminals Behind The Badge

The List: What a Jail Phone Vendor Actually Keeps

Six categories, one account, nineteen months — the complete records demand filed against a jail phone vendor.

The request dated 9 December 2025 went to NCIC Correctional Services for the complete communication and billing history attached to one account at the Williamson County Jail, covering 6 January 2024 through 15 July 2025. The requestor is both the account holder and the person who was incarcerated — and the identity section says so plainly: no third party, and no privacy restriction claimed on his own behalf. A request for records is only as good as the categories it names — this one names six, and read together they describe what a modern jail telephone contract actually produces.

One — call detail records:

Every outbound and inbound call, with date-and-time stamps, duration, successful and failed connection attempts, destination numbers, and recording file identifiers where a recording exists. The recording identifier is the part with teeth: if a recording exists and carries an identifier, the audio is a separately retrievable object — and if the identifier points at a file that no longer exists, retention policy has become a fact, not a claim.

Two — video visitation logs:

Dates, times, durations, session identifiers, connection history including failed and terminated sessions. A failed video visit is not a neutral event — it is a missed hour with a child, and if a log shows a pattern of terminations at the same point, that pattern demonstrates defective service sold as working.

Three — messaging activity:

All tablet and kiosk messages sent or received — with timestamps, message-credit billing logs, and blocked or failed delivery logs. Messaging is where a jail communication platform touches the most sensitive content at the lowest cost to the vendor — and where the billing model is most opaque: credits are purchased, messages are charged, and delivery is not guaranteed.

Four — account financial activity:

Charges, deposits, fees, adjustments, transaction timestamps, linked phone numbers, prepaid balance history, and any charges drawn against an inmate trust fund. A consumer-protection regulator would read this category first — it is the only one that can show whether money left an account without a corresponding service.

Five — device and system metadata:

IP and device identifiers associated with messages, login and access logs, account creation metadata. Most people who use a jail messaging app are not incarcerated — they are mothers, siblings, attorneys and friends installing an application on their own phones because it is the only way to stay in contact. The metadata request reaches their devices — precisely why a vendor may resist producing it.

Six — blocks, restrictions and flags:

Any permanent or temporary block on the account, the reason code, the date and time scope of each block, and any system-level restriction imposed by jail staff or by the vendor. The smallest category and the most revealing — a block is a decision made by someone, and a reason code identifies who made it and on what authority.

“The list is the point. Six categories, one account, nineteen months.”

Formats — why the format is part of the request:

The request specifies PDF for logs, CSV or Excel for datasets, and MP3 or MP4 for audio and video identifiers. A PDF is a picture of a table and cannot be sorted; a CSV can. A vendor that produces a scanned summary in place of a dataset has not produced the record — the format clause is how that becomes visible instead of arguable.

What the Record Shows:

If the records do not exist, the answer is a sentence saying so — and that sentence is itself a finding about retention. Production, partial production, denial with a reason, or silence — each is a documented outcome, each can be tested. The interesting question is not whether the list can be answered — any county that wants to argue its jail telephone system is ordinary should be able to produce all six without difficulty. It is what happens when it is asked.