Most workplace violence prevention plans are written by HR, reviewed by legal, filed in a shared drive, and never tested against the one place violence actually enters a building: the front door.
That gap is not theoretical. California's SB 553 has required a written Workplace Violence Prevention Plan from nearly every employer in the state since July 1, 2024. Other states are drafting similar language. And federally, OSHA has never needed a specific workplace violence standard to cite an employer — the General Duty Clause, Section 5(a)(1), already obligates you to keep the workplace free from recognized hazards, and inspectors have used it for workplace violence citations for years.
So the plan gets written. Roles get assigned. Training gets scheduled. And then a contractor walks past an unattended lobby desk at 7:40 a.m., signs a paper sheet with an unreadable name, and disappears into the building for six hours.
Here is the uncomfortable part: if something happens, the paper sheet is your evidence. It is what you hand to investigators. It is what your insurer reads. It is what opposing counsel enlarges and puts on a screen.
What the regulations actually ask you to prove
Read a workplace violence prevention requirement closely and you will notice that most of it is not about preventing a single dramatic event. It is about documented, repeatable process. Under SB 553's framework, the plan has to cover:
- Named responsibility. Specific people who own implementation, not "management."
- Employee involvement. Procedures for staff to participate in developing and running the plan.
- Hazard identification and correction. An ongoing process — inspections, assessments after incidents, correction of what you find. Records kept a minimum of five years.
- Reporting without retaliation. A confidential, accessible way for anyone to raise a concern.
- Incident response and investigation. How you investigate, what you document. The violent incident log and investigation records also carry a five-year retention minimum.
- Training. Delivered to all employees, with training records kept at least one year.
- Coordination with other employers. When you share a space with tenants, staffing agencies, or contractors, someone has to reconcile the plans.
- Emergency response integration. Violence procedures folded into the broader safety program, not bolted on.
Notice how much of that is a records problem. Five years of hazard correction records. Five years of incident logs. A year of training records. Coordination documentation across employers sharing a building.
Now ask what your visitor log contributes to any of it.
The four types of workplace violence, mapped to your lobby
Cal/OSHA and NIOSH both use a four-type classification, and it is worth walking through because each type has a different relationship to the entrance.
Type 1 — Criminal intent. The person has no legitimate relationship to your organization. Robbery, trespass, targeted attack. This is the type everyone pictures, and it is the type a controlled entry point most directly addresses. An unstaffed side door and a propped-open loading entrance are Type 1 hazards. So is a lobby where nobody can tell a stranger from a vendor.
Type 2 — Customer, client, patient, or visitor. Someone with a legitimate reason to be there who escalates. Healthcare sees the most of this by a wide margin, but so do schools during custody disputes, government service counters, and property management offices during eviction season. This is the type where a visitor management system earns its keep, because the person is in your system — and the system either knows their history or it does not.
Type 3 — Worker on worker. Current or former employees. Offboarding is a security event, not just an HR event. The former employee who still has a working badge is the most common version of this failure.
Type 4 — Personal relationship. Domestic violence following someone to work. The threatening party has no relationship to the employer at all, only to one employee. This is the scenario where a discreet watchlist entry and a front-desk alert matter more than a locked door, because the employee often knows exactly who to watch for and has no safe way to tell the building.
A paper sign-in sheet addresses exactly none of these. It records that someone wrote something at approximately some time.
Five places the plan usually breaks at the entrance
1. You cannot produce a visitor list for a specific time window
An investigator asks who was in the building between 2:00 and 4:00 p.m. on a Tuesday eleven months ago. With a paper log, that is a search through boxes, assuming the boxes still exist and the pages were not thrown out at the end of the quarter. With a digital record, it is a query.
The five-year retention expectation in the regulations is not a suggestion. If your visitor records have a shorter practical lifespan than your legal obligation, the gap is the finding.
2. Nobody checks out
Check-in is the part everyone designs. Check-out is the part that determines whether your evacuation roster is real. If a third of visitors never sign out, your occupancy count is fiction — and "knowing the count and location of visitors at all times" is exactly what emergency protocols assume you can do.
Automatic end-of-day check-out with a flag on anyone who never scanned out is a small feature that converts a decorative number into a usable one.
3. There is no watchlist, or the watchlist lives in one person's head
This is the single biggest gap between a plan and an implementation. An employee reports a threat from an ex-partner. The plan says the report is documented and acted on. In practice it gets told to the front desk supervisor, who tells the day receptionist, who is out sick on the day it matters.
A watchlist inside the check-in system changes that into an automatic alert: the name or ID hits at the kiosk, the badge does not print, and a designated responder is notified before the visitor reaches the elevator. That is the difference between a plan that exists and a control that operates.
Two things to get right when you build one. Restrict who can add and read entries — a watchlist is sensitive personal data and it deserves access controls and an audit trail. And define a removal process, because an entry with no expiration date becomes an indefinite private blacklist nobody reviews.
4. Contractors and vendors are governed by a different, worse process
The coordination-with-other-employers requirement exists because shared spaces create orphaned responsibility. The staffing agency assumes the building screens. The building assumes the agency screens. Meanwhile the temp worker has a badge, a code, and no record in either plan.
Put every non-employee through one entry process, regardless of who signs their paycheck. Recurring vendors get a pre-registration profile so repeat entries are fast, not skipped.
5. Your incident log and your access record are separate systems
When an incident is investigated, the two questions are always: what happened, and who was present. If your violent incident log is a document and your entry record is a different system that nobody exports, reconstructing a timeline takes days and produces gaps.
You do not need one monolithic platform. You need an entry record you can export, timestamp-accurate, and attach to the incident file.
Turning the plan into controls that actually run
Here is a practical sequence that maps plan requirements onto the entrance, in the order that gets results fastest.
Step 1 — Inventory every way a person can enter. Walk the building at the times it is most permeable: shift change, early morning deliveries, evening cleaning, event days. Count the doors that are propped, badge-free, or watched by nobody. This walk is the hazard identification your plan requires — document it, date it, and keep it.
Step 2 — Make one controlled entry point the default. Everything else becomes exit-only or badge-only. This is unglamorous and it does more than any software purchase.
Step 3 — Replace paper with a digital record at that point. Government ID capture where appropriate, host notification on arrival, printed badge with photo, timestamp on both ends. The goal is a record you can query, not a nicer-looking sheet.
Step 4 — Enable pre-registration for anything scheduled. Screening that happens the day before is screening that actually happens. Same-day walk-ups get more scrutiny, not less, because they are now the exception.
Step 5 — Build the watchlist and route the alerts. Decide now: who receives the alert, on what device, and what they do in the next sixty seconds. An alert with no defined responder is a notification, not a control.
Step 6 — Define access by visitor type. A job candidate, a plumbing contractor, and a parent picking up a student should not receive the same badge with the same implied access. Different badge colors, different escort rules, different floors.
Step 7 — Test the evacuation roster. During the next drill, pull the live visitor list and physically reconcile it at the assembly point. Whatever gap you find is your real number. Fix it, then retest.
Step 8 — Close the loop on offboarding. Termination triggers badge revocation the same day, and the former employee's name goes into the visitor system with an escort-required or alert flag if the separation warrants it.
Step 9 — Train to the specific scenario. Generic awareness training does not help a receptionist facing an escalating visitor. Role-play the three situations most likely at your site, including the one where the correct answer is to press the duress button and say nothing.
Step 10 — Review quarterly, with records. Pull entry data, look at after-hours entries, unclosed visits, watchlist hits, and badge anomalies. Document what you found and what you changed. That document is your hazard-correction record, and it is the artifact that turns a binder into a program.
What "reasonable" looks like after an incident
Liability after a workplace violence incident rarely turns on whether you prevented it. It turns on whether your response was reasonable given what you knew and what similar organizations do.
Reasonable, in 2026, increasingly means: you had a written plan, you identified entry as a hazard, you controlled the entrance, you kept a queryable record of who was inside, you acted on reported threats through a documented mechanism, and you can show the last four quarters of review.
Unreasonable looks like a clipboard, a propped door, and a threat report that was mentioned verbally to someone who no longer works there.
The distance between those two is a few weeks of deliberate work, not a capital project.
Start with the question you cannot currently answer
Pick the hardest one and try to answer it today, without preparation:
Who was in this building at 3:15 p.m. on the second Tuesday of last month, and when did each of them leave?
If the answer takes more than two minutes, your workplace violence prevention plan has a front door problem — and the fix is a documentation problem before it is a security-technology problem.
KyberAccess provides visitor management for schools, healthcare facilities, and multi-site organizations: digital check-in, ID verification, watchlist alerts, host notifications, and exportable entry records built for audit and emergency response. This article is general information, not legal advice — confirm your obligations with counsel, since requirements vary by state and industry.