Understand MRC triggers and pre-population¶
Two separate mechanisms, often confused¶
1. Pre-population: does a field type-ahead, or do you free-type it?¶
Per FDD Common 10.1.2 and the "MRC Employer/Jobsite/Contract pre-population" cluster
(task-management/cases/closed/MRC-CLUSTERS.md), whether the Employer/Jobsite/Contract fields
on a ticket type-ahead and pre-populate from existing PRIZM data, or must be free-typed, is
governed entirely by the Category selected on the ticket — not a bug tied to a specific
jobsite or employer.
The rule: any Category referencing an existing Employer/Jobsite/Contract type-aheads and pre-populates from PRIZM. Any "-NEW" Category (new Employer, new Jobsite, new Contract) is always free-text, by design — there's nothing in PRIZM yet to look up, since the record doesn't exist until the ticket's action creates it.
Three independent tickets all traced back to this one rule:
- "Change of management" fields not prepopulating traced to Category simply not being selected, or set to the wrong value.
- Employer not updating to match a newly selected jobsite during a job change was a confirmed, separately fixed bug — not the same mechanism as the Category rule.
- "New Employer type-ahead not working" was not a bug at all — "-NEW" categories never type-ahead by design.
2. Triggers: which tickets does PRIZM create automatically?¶
Per FDD Common 10.1.3 (MRC Assignment Triggers), a defined set of PRIZM events causes the system to generate an MRC ticket automatically, with Category, data fields, and assignees all pre-determined by the event type. Automatic creation runs as a nightly server job, not in real time — the job reviews the day's change-history records each night and creates the corresponding tickets. That job does not retroactively catch up changes made while it was known to be down.
Events that do trigger an automatic ticket include (per the FDD's Event Triggers section):
- New member record created (with an active-type status such as Normal Member, Hire/No app, Partial Member, Permit, or Transfer In)
- Change in Member Status (out of Pending E-Board) or Member Type
- Change in Member Trade Code, Job (employer/jobsite), or Local
- Member FedPAC/COPE participation change
- New Jobsite, Jobsite change of management
- New Employer
- New Contract, Contract status change/renewal, Contract contacts
Each triggered ticket defaults to: Assignment Due Date 3 business days out, "system"/"PRIZM MRC" as the creating/assigning user, next sequential ticket number, and Open status. The user whose action caused the trigger is not assigned the resulting ticket themselves (they're assumed to have already acted), but their Fund Group is still notified.
Categories that are explicitly manual-only — the FDD lists these outright as "Issues That are Not Triggered by PRIZM Events" — include Temporary Worker requests, Member H&W Eligibility, general demographic/life-event changes (name/address), Member Packet Requests, Member COBRA Eligibility, H&W Eligible-not-Enrolled, and general (non-trigger-worthy) Jobsite/Employer/Contract changes or contract term changes. These are ad hoc by nature or would otherwise generate too much clutter if auto-triggered, so the FDD leaves them to a human's judgment to create a ticket when appropriate.
Things to watch out for¶
Give the nightly batch time — don't conclude the trigger is broken within minutes of a change
Trigger-based MRC tickets are batch-generated overnight, not instant, and the batch does not backfill missed days. A Deceased status change not creating an MRC traced to two causes at once: Member Status and Member Type are independent fields with no auto-cascade between them, and the nightly trigger batch doesn't run retroactively for changes made while it was down. (IUOE-2523)
A terminated work record correctly generates an MRC roughly 15-20 minutes later
A terminated work record not creating an MRC ticket for H&W was a real, fixed bug — shipped
in r2.27.01 and UAT-validated with a fresh repro (Ramirez, Ricardo) showing the ticket
appearing about 15-20 minutes after the triggering change, correctly addressed to
H&W/ETF/Membership/Business Agent. See e2e/uat-validation/IUOE-2748/FINDINGS.md for the
full repro. If a ticket you expect from a work-record change hasn't shown up yet, also check
the Fund dropdown scoping described in
Find tickets across funds and assignments —
that's what actually cost ~15 minutes of false "nothing created yet" checks on this same
case. (IUOE-2748)
'-NEW' categories are always free-text — that's not a missing type-ahead
Before filing a ticket that a New Employer/Jobsite/Contract field "isn't typing ahead," check whether the Category selected is one of the "-NEW" categories. If so, free-text entry is the documented, correct behavior, since the record being described doesn't exist in PRIZM yet.
A reference-table entry not disappearing after being marked inactive is unconfirmed
One open report (an MRC Reference Table Entry not disappearing after being made inactive) is still only a Phase 1 stub — possibly a related trigger/caching mechanism, but not yet confirmed. Don't assume it's the same root cause as the nightly-batch behavior above without further investigation.
See Log and track an MRC ticket for how a manually-entered ticket's required fields and Assignment History work once either path — manual entry or an overnight trigger — has created it.
Who can do this¶
| Role | Access |
|---|---|
| Super User | Add/Update |
| Business Agent | Add/Update |
| Management | Add/Update |
| Membership Staff | Add/Update |
| Accounting | Add/Update |
| Contract | Add/Update |
| H&W Staff | Add/Update |
| ETF Staff | Add/Update |
All 8 roles can see and act on MRC tickets regardless of whether the ticket was entered by hand or generated automatically — there's no separate permission for "trigger-created" tickets. What differs isn't who can see them, it's understanding when PRIZM will create a ticket for you versus when someone has to enter one manually. Getting that mental model wrong is the single biggest source of "this is broken" reports against MRC that turn out to be expected behavior.