On September 1, Nebraska launched Nebraska Navigator, an AI assistant on the redesigned Nebraska.gov. Governor Jim Pillen announced it the next day. Residents can ask about state services without searching across agency websites. GovTech reports that Tyler Technologies developed it with the state. The Nebraska State Records Board pays for it with transaction fees from online government payments.
It grew out of a pilot at the Department of Motor Vehicles called Ask the DMV. The pilot’s results are the case for going statewide, and they are worth reading carefully.
Tyler announced that the assistant answered more than 88,000 questions from more than 38,000 users between May and August. The governor’s office put the figure at more than 188,000 questions since the pilot began in December 2025. Different counting windows may explain the gap. Neither report says.
The governor’s office reports a 20 percent drop in call volume. StateScoop, citing the DMV’s deputy director, describes it as a drop in some areas.
Before the statewide launch, state teams tested the assistant with more than 5,000 questions and reported a 95 percent success rate. The assistant is in public beta into October. Officials say resident feedback will guide when individual agencies adopt it.
Nebraska expanded on pilot numbers. Then it staged the expansion, which is the more useful decision.
What the numbers show, and what they don’t
Usage figures show that people use the tool. About 19 percent of interactions came between midnight and 6 a.m., which is demand a call center doesn’t serve. A drop in calls shows work moving from phones to the website.
None of that shows the answers are right. The 95 percent figure comes closest, but the public reporting doesn’t define success or say who wrote the test questions. No independent evaluation has been reported.
Compare Denver, where the city council declined to renew its chatbot contract despite a reported 90 percent satisfaction rate. Satisfaction and usage answer one question. The decision to expand asks another.
Questions to ask before you expand a pilot
- Who reported each number: the vendor, the agency, or someone independent?
- What window does each figure cover? Totals from different periods can’t be compared.
- Is the drop in calls measured across the whole contact center or in selected queues?
- How is success defined in accuracy testing, and who wrote the test questions?
- What happens when the assistant doesn’t know the answer? Is there a handoff to a person? Nebraska’s public reporting doesn’t say.
- Does each new agency get its own testing before going live, or does it inherit the DMV’s results?
A pilot proves people will use the tool. Only testing shows whether it’s right.
Sources