← Back to library
OperationsOpen4 min read · 20 min to run

California Just Banned AI-Only Firings: The 20-Minute Inventory of Every Tool That Already Scores Your People

SB 947 was signed September 30 and takes effect July 1, 2027. If software ever rates, ranks, or flags an employee and that output drives a write-up or a termination, a person has to confirm it first. Most owners cannot name which of their tools do that. This finds them.


What happened: On September 30, 2026, California's governor signed SB 947, the No Robo Bosses Act. Starting July 1, 2027, a California employer cannot discipline or fire an employee based solely on an automated decision system. When software is the primary basis for the decision, a human reviewer has to corroborate it with other evidence, such as the personnel file or a manager's own evaluation. The employee gets written notice that software was involved and can ask what data about them it used. Penalty: $500 per violation, enforced by the Labor Commissioner. There is no small-employer exemption. It covers employees, not contractors.

The definition is wider than you think. The law covers any computational process built on machine learning, statistics, or AI that produces a score, classification, or recommendation used to assist or replace a human judgment about an employee. That reaches further than a chatbot. A timeclock app that flags "excessive late punches." A dispatch tool that ranks techs by revenue per hour. A call-recording tool that grades customer service calls. A GPS platform that scores driving. If one of those outputs ever ended up in a write-up, it is in scope.

The so-what if you are outside California: Several states have similar bills in motion, and the human-confirmation step is the right habit anyway. A score that nobody checked is how a good employee gets written up for a timeclock glitch.

The move: inventory the tools, then add the human step. Paste this into Claude or ChatGPT:

I run a [type of business] with [number] employees in [state]. Below is every piece of software my company uses that touches employees: scheduling, timekeeping, payroll, dispatch, GPS, call recording, performance dashboards, hiring tools, and any AI assistant a manager uses to write reviews.

[PASTE LIST, one tool per line, with what it does in your words]

For each tool:
1. Does it produce a score, rank, flag, or recommendation about a person? Answer yes, no, or unsure.
2. If yes, what other evidence a manager would need to confirm that output before using it in a write-up, a schedule cut, or a termination.
3. One plain sentence a manager could put in a write-up stating that the tool was used and what it showed.

Then list every tool marked unsure so I can ask the vendor.

FORMAT: One table, then the unsure list. Do not give legal advice. Flag any line where I should ask an employment attorney.

Then:

  • For every "yes" tool, write one line in your discipline procedure: a manager confirms the output against a second source before acting on it.
  • Add a line to your write-up template for the tool used and what it showed.
  • Ask each "unsure" vendor one question: does your product score, rank, or flag individual employees?

Keep it safe: Describe the tools. Never paste employee names, pay, or any individual's data into the chat. The effective date and the notice details should be confirmed with your employment attorney before July.

What it's worth: Twenty minutes now. One contested termination costs more than any software on the list.