On July 1, California became one of the first governments in the world to roll out a generative AI assistant to every state agency. The platform is called Poppy. It runs 10 models. It pulls information only from official CA.gov sources. And it was built not by a vendor, not by a consulting firm, but by state employees.
That last part is the detail most people will skip past. It is the only one that matters.
The Pilot That Actually Worked
Poppy started as a pilot in September 2025. The California Department of Technology put it in the hands of 2,800 employees across 67 state departments. Not a handpicked innovation team. Not a digital-first agency. Sixty-seven departments, which means DMV clerks and environmental regulators and procurement officers all testing the same platform.
The pilot ran for nine months. Employees gave feedback. The CDT iterated. And on July 1, 2026, it went statewide.
Compare that to the average enterprise AI pilot. Most companies pick one team, run a 90-day test, produce a report, and then spend the next year debating whether to expand. California tested across 67 departments simultaneously and shipped to production in under a year.
The platform itself is boring on purpose
Poppy is not a custom model. It is not a proprietary system. It is a platform that consolidates 10 commercially available AI models into a single interface, configured for the specific needs of state government work.
Employees can draft documents, analyze data sets, and research policy. They can compare outputs from multiple models side by side in a single view. The interface includes pre-built query templates for common state business tasks so that someone who has never touched AI before can get useful results on day one.
Here is what Poppy does not do: it does not send data outside California’s trusted environment. It does not use employee inputs to train the underlying models. The CDT’s guarantee is four words: “no model training.” Every interaction stays inside the state network.
This is not sophisticated. That is the point.
The Design Decision That Explains Everything
Most enterprise AI deployments fail at adoption, not capability. The model can do the work. The employees do not use it. They do not trust it. They do not understand it. Or they tried it once, got a bad result, and never came back.
California designed around that exact failure mode. Pre-built templates mean nobody has to learn prompting. Trusted CA.gov sources mean employees do not have to evaluate whether the output is pulling from reliable data. The “no model training” guarantee means the privacy conversation is settled before it starts. And vendor-agnostic model access means no one is locked into a single provider’s interface or pricing.
Every one of those decisions is a simplicity decision. Not a technology decision.
So what would yours look like
The question business leaders should be asking is not “which AI model should we buy.” It is “what would our version of Poppy look like.”
Start with the work your people actually do every day. Not the aspirational use cases from the vendor deck. The real tasks. The document drafts, the data pulls, the policy lookups, the repetitive analysis that eats hours every week.
Then build the simplest possible interface that connects AI to those tasks. Pre-built templates. Trusted data sources. Clear privacy boundaries. Something a new hire can use in their first week without asking IT for help.
California did not pick the best model. They did not hire an army of AI consultants. They did not wait for agents to mature. They took 10 existing models, wrapped them in an interface their employees could actually use, restricted the data environment to sources they already trusted, and shipped it.
2,800 pilot users across 67 departments. Statewide rollout in under a year. Built by the people who will use it.
Meanwhile, most large companies are still circling. New vendor eval. New pilot cohort. New steering committee. A state government — the DMV people — shipped production AI to every agency while the Fortune 500 was still scheduling alignment meetings.
The bottleneck was never the model. It was the willingness to make the tool simple enough that people would actually use it. California got that right. If your organization hasn’t, the uncomfortable truth is that a government bureaucracy just lapped you.