California's FAST Act, Nearly Two Years In: AB 1228 Is an Automation Accelerant

It is Monday morning in the newsroom and the only thing on my screen worth reading twice is a calendar entry. April 1 is three weeks out, which means California’s $20 fast-food minimum wage — the operative floor under Assembly Bill 1228 — hits its two-year anniversary in the same month operators close their Q1 books. The 2025 disclosures are legible enough now to put on the page. The wage floor did not cause the voice-AI and kiosk push in California QSR. It accelerated it. Interpretation, not a finding.

The formal academic write-up is days away. A forthcoming UC Santa Cruz report from economics lecturer Stephen Owen, due later this month, will quantify the FAST Act’s second-year labor effects across multiple California QSR cohorts. The numbers operators have been hearing in private since the fall get a citable source for the first time. I have not read the report. I am reading the pre-publication advisory.

What we know before the report drops

The April 1, 2024 effective date is settled history. The Fast Food Council the bill created has spent the interval on petitions and a pending vote on a 2026 increase. Operators have spent it recalibrating crew schedules, kiosk procurement, and voice-agent pilots against a fully-loaded California labor cost that runs meaningfully above the federal QSR comp.

Two things are knowable today without the Santa Cruz numbers.

One. Every multi-unit California operator I have spoken to since the fall frames automation procurement in California with a different urgency than in their other markets. Directional, not quantified. Vendor decks landing in California in 2025 lead with payback-period math indexed to the $20 line. The same vendors pitching into Texas or Florida lead with throughput. The wage line is the wedge.

Two. The deployments are disclosed. Burger King’s Patty assistant — the OpenAI-backed crew-and-manager surface in 500 U.S. stores per the Nation’s Restaurant News piece — is national, but the labor contract bites hardest where crew cost is highest. California is the marginal market that decides whether the rollout gate clears. McDonald’s voice-ordering programme, which Leila is writing up in long form for The Operator, runs the same calculus from the other end of the labor stack. Both deployments would exist without AB 1228. Neither would have moved this fast without it.

The Owen report, as I read the advisory

The framing on the UC Santa Cruz advisory page is explicit. It is a labor-economics paper measuring employment, hours, and shift-work changes across California QSR cohorts before and after the wage line took effect. It is not a causal-inference paper trying to isolate the AI-deployment variable from the wage variable. Read it that way and you will read it correctly.

The headlines operators are bracing for track two cohorts the advisory has named. First, Burger King locations under at least one coastal-California franchise owner — a previewed decline of more than 21 percent in shift work from October 2023 to October 2024. Second, 18 McDonald’s franchise locations in the Central Valley — nearly 12 percent decline in total labor hours across the twelve-month windows from April 2023 through March 2025, equivalent to roughly 62 full-time-equivalent jobs. Previews from the advisory, not the paper. I am holding them at arm’s length.

The question the report will not answer is the one operators most want answered: how much of the decline is kiosk-and-voice substitution, how much is schedule compression, how much is reduced hours, and how much is closures. The advisory does not split it. Dan Walters has a forthcoming column on the FAST Act’s measured effects that will do part of that interpretive work. Walters writes opinion, not findings.

The accelerant frame

The wage floor is not the cause of California QSR automation in 2026. The cause is the one driving QSR automation everywhere — voice agents and kiosks crossed the productisation line, the model cost curve broke right, and integrators stitched the surfaces into POS and KDS stacks well enough to ship. The wage floor is the catalyst. It accelerated a deployment timeline that would have unfolded anyway, by an amount operators in lower-cost states will quantify two to three years from now.

That distinction changes the pitch. A vendor selling kiosk-and-voice into California as the answer to AB 1228 is selling a story that will not survive the Owen report. The labor-hour declines pre-date the heaviest voice-AI deployments by a year or more. The wage floor moved the timeline. The technology readiness moved the deployment. Operators who buy the answer to AB 1228 pitch will under-budget the engineering work after the kiosk goes live.

It also changes the budget conversation. AB 1228 is a forcing function for an automation roadmap that was already coming. The right question is not what do I buy to absorb the $20 floor, but what does my California operating model look like in 2028, and what do I need to procure now to get there. Eitan’s Four Margins frame, running in long form later this spring, sits one level above this question. This piece sits at the news cycle. The frame sits at the strategy.

What to watch between now and the print

One. The Owen report itself, due later this month from UC Santa Cruz. Read the methodology section first — cohort selection and windowing do the analytic work, and the headline percentages are intelligible only against that scaffolding.

Two. The Walters CalMatters column, due around the same window. Commentary, not finding. The political-economy frame will be useful. The numbers will not be his.

Three. The Fast Food Council’s 2026 wage-adjustment vote. Clear an increase on top of the $20 floor and the labor curve steepens, the procurement curve with it.

The piece every voice-AI vendor and California QSR operator should be ready to file by April 1 is a quantified one: what did the FAST Act do to the labor line, what did the technology do to the deployment line, which moved faster. Today I do not have those numbers in writing. In two weeks I expect to.

— Hana edits the newsroom for TableTransfers. Tips: [email protected].