Keyword Monitoring is the watchlist engine sitting in front of every inmate communication stream — phone, mail, electronic messaging. Analysts curate keyword sets; the system raises alerts; a simple lifecycle keeps the queue moving. It’s the force multiplier that lets a small intel team cover a large population without missing the calls and letters that count.
What it does
- Lets administrators maintain keyword sets organized by category — contraband (drugs, weapons), gang reference, violence indicator, escape language, PREA-related — with per-keyword active state, owner, and last-hit metadata.
- Generates an alert each time a keyword is detected, tagged with channel, source record, inmate id, severity, and the matched context (a window of words around the hit).
- Provides a New → Reviewed → Escalated → Dismissed lifecycle so the queue actually drains, supervisors see what was escalated to them, and the alert never goes silently stale.
- Surfaces keyword effectiveness — hits per keyword per week, false-positive rate from analyst dismissals — so the watchlist improves with every review cycle.
How it works
Detection runs against transcripts (Call Monitoring), scanned mail (ScanMail), and messaging records in the IPS layer. Hits create alert rows in the JMS database with the matched phrase, the surrounding context, and the originating record id — so an analyst can jump from the alert straight to the source call or letter, and from there to the inmate’s full file.
Keyword changes take effect on the next detection cycle. Adding a keyword does not retroactively scan archived calls by default — that’s a separate (and auditable) reprocessing action, gated behind supervisor permission.
Data captured
- Keyword string, category, active flag, owner, creation timestamp
- Per-alert: inmate id, channel, source record id, matched context, severity, current status
- Reviewer id, decision timestamp (UTC, displayed facility-local), escalation chain, dismissal reason
- Keyword effectiveness rollups for tuning
Compliance & retention
Keyword Monitoring is an HSGP "Secure Communications" control. The alert audit trail — who saw an alert, when, and what they decided — is preserved per facility records policy. Any alert that becomes evidence in a Threat Assessment is stamped with its case id and inherits the case-linked retention window. Privileged channels are excluded from keyword scanning by policy, with the exclusion logged for audit.
We’ve been manually flagging calls for years and this kind of automated keyword routing would save our analysts so much time. Quick question — can the watchlists be customized per facility or are they shared system-wide? That distinction matters a lot for how we’d want to configure alerts across our two housing units.
We implemented something similar last year and the keyword alert lifecycle was honestly the piece we didn’t know we needed — our old system just flagged everything with no way to track who reviewed it or what happened next. Does your platform allow supervisors to customize alert routing by housing unit, or is it facility-wide only?
We rolled out keyword monitoring on inmate calls about eight months ago and the alert routing piece alone saved our analysts hours of manual review every week. One thing I’d ask — does the system allow different watchlists per housing unit, or is it one global list across the whole facility? That flexibility would be a real game-changer for us.
We’ve been using keyword flagging on calls for about eight months now and the alert lifecycle piece is what really sold our command staff — knowing something didn’t just get flagged and forgotten makes a huge difference operationally. Does your system allow supervisors to add custom keywords on the fly without involving IT, or does that still require a backend change? That’s been our biggest friction point with our current setup.
We implemented something similar last year and the lifecycle tracking piece is what sold our command staff — being able to show that every flagged item got reviewed and documented was huge during our last audit. Quick question though: can analysts add custom keywords on the fly without needing IT involvement, or does that require an admin config change?
We rolled out keyword monitoring on inmate calls about eight months ago and the alert lifecycle piece is what sold our command staff — having that New/Reviewed/Escalated workflow keeps things from falling through the cracks during shift changes. One question though: does your system allow different watchlists per housing unit, or is it one global list for the whole facility?