Home / Services / AI governance and SB 947

AI governance and SB 947 readiness.Central Valley

Your staff are already using AI tools at work. The question is whether the rules match reality: which tools are approved, which data may go into which tools, and who decides when a tool’s output feeds an employment decision. California has now answered part of that question by law — and the deadlines are set. SB 947, the No Robo Bosses Act, becomes operative July 1, 2027. Three more workplace AI laws take effect January 1, 2027.

We give Central Valley employers a clear answer to what their people may do with AI tools, and with which data — aligned to the NIST AI Risk Management Framework, written for how the organization actually works. Every engagement starts with a first look, at no charge.

The work is reviewed and signed by Brian Olson, CMMC Registered Practitioner.

When it comes up

  • Staff use AI tools with no policy, and nobody knows which tools are in use or what data goes into them.

  • You screen applicants or make employment decisions with AI — California’s FEHA automated-decision-system regulations have covered that since October 1, 2025.

  • A tool scores or flags employees for discipline or termination — SB 947 applies from July 1, 2027, regardless of company size.

  • Your work involves controlled or export-controlled data, and consumer AI tools are off-limits for it.

  • A customer asks how you govern AI, or you want certification to the AI management standard (ISO 42001).

What you receive

  • An AI readiness check

    Which tools are in use, and for what, including tools adopted without IT or management approval.

  • A written AI use policy

    Which data may go into which tools, who decides, and what the human step is when a tool’s output feeds a decision.

  • An approved-tools list

    Kept current as the tools change, so the rules still match what your people use.

  • A written read of which California AI rules land on your business and which do not

    That second half matters: several heavily publicized AI laws reach only large developers, and most businesses are outside them. You should spend your time on the ones that reach you.

  • For public agencies: the paperwork for a generative AI risk review

    Prepared alongside the firm that already holds the agency’s contract. California state agencies must complete a generative AI risk assessment before buying a tool.

The SB 947 deadline

From July 1, 2027, a California employer may not rely solely on an automated decision system for a discipline or termination decision. If the employer primarily relies on the system’s output, a person must corroborate the decision from the underlying data, and the employee must receive a written notice. The penalty is $500 per violation. The work — inventory, procedure, notice, manager training — should be done before July 2027, not started then. Read more: SB 947: what it means for a small California employer

Common questions

Straight answers.

What people ask about AI governance and SB 947.

We only use ChatGPT and Copilot. Do we need a policy?

If anyone uses AI with customer files, pricing, personnel records, or controlled data, yes. The policy sorts data into what may go into which tools — that sorting is the control, and it matters more than which brand of tool is involved.

Can we use AI with controlled or export-controlled data?

Not in consumer AI tools. If your contract carries DFARS 252.204-7012, a cloud service handling controlled information must meet the FedRAMP Moderate baseline or its equivalent. Export-controlled data has its own limits. Your AI use policy sorts this out. This is not legal advice; for export questions, ask export counsel.

Does this replace legal advice?

No. We document your position and write your policies. How a specific law applies to a specific tool or decision is a question for employment counsel, and we say so in the report.

What does it cost?

It depends on scope, so we don’t list prices. The scope is in writing before any work starts. The first look costs nothing.

First look

Start with a first look.

A short review of your AI use and where you stand, and a short report signed by a person. No charge, no obligation. A person gets back to you within one business day.