An analysis of the federal government’s own web analytics data reveals that AI tools have become a measurable gateway to public services, 43.5% of citizens access government sites by phone, and nearly 1% are still using a browser that was discontinued in 2022.
Analysis by VWO.com | All data sourced from Analytics.usa.gov (U.S. federal government official analytics) | Data period: Dec 28, 2025 – Jan 26, 2026
Something unusual is happening in the data behind America’s government websites. Buried in the federal government’s own publicly available analytics dashboard, a signal is emerging that would have been unthinkable two years ago: ChatGPT is now a measurable source of referral traffic to federal services, sending as many visitors to government websites as Google’s paid search ads.
An analysis of data from Analytics.usa.gov, the official U.S. government web analytics platform operated by the Digital Analytics Program (DAP), conducted by VWO.com, a digital experience optimization platform, reveals that ChatGPT accounts for 0.7% of referral traffic to federal government websites, matching Google’s paid search traffic at 0.7% and exceeding referrals from major platforms like eBay (0.5%). The data covers the 30-day period from December 28, 2025 through January 26, 2026.
But the AI finding is just one thread in a broader picture of how Americans navigate their digital relationship with the federal government. The same data shows that 43.5% of citizens are accessing government services on their phones, nearly half of all visitors still depend on search engines to find federal services, and almost 1% are doing so on a browser that was officially discontinued in 2022.
AI as a Government Services Gateway
The federal analytics data breaks down referral traffic by source. Among the smaller but notable referral channels, ChatGPT’s presence is striking, not for its absolute size, but for what it represents.
| Traffic Source | % of Total Users | Traffic Type |
| ChatGPT.com / Referral | 0.7% | AI-assisted referral |
| Google / Search (Paid) | 0.7% | Paid search |
| eBay.com / Referral | 0.5% | E-commerce referral |
| cn.bing.com / Referral | 0.4% | International referral |
Source: Analytics.usa.gov, U.S. federal government official web analytics. Data period: Dec 28, 2025 – Jan 26, 2026.
That 0.7% figure means that Americans are now using AI assistants to navigate to government services at the same rate as those who click on Google’s paid advertisements. While still a small fraction of total traffic, it establishes AI tools as a measurable channel alongside traditional referral sources. The trajectory matters: this is a channel that effectively did not exist in government analytics data two years ago.
How Americans Actually Find Government Services
The analytics data reveals a population split almost evenly between two navigation behaviors. According to the federal data, 48.6% of visitors navigate directly to government websites (typing URLs, using bookmarks, or clicking links from emails and documents), while 45.6% arrive through search engines. The remaining 5.8% come through referral links and other channels.
| Traffic Source | % of Total Users | Rank |
| (Direct) / (None) | 48.6% | 1 |
| Google / Organic | 40.9% | 2 |
| Bing / Organic | 3.4% | 3 |
| Other | 2.9% | 4 |
| Secure.ssa.gov / Referral | 1.2% | 5 |
| Chatgpt.com / Referral | 0.7% | 6 |
| Google / Search | 0.7% | 7 |
| Yahoo / Organic | 0.6% | 8 |
| Ebay.com / Referral | 0.5% | 9 |
| Cn.bing.com / Referral | 0.4% | 10 |
Source: Analytics.usa.gov. “Direct” includes typed URLs, bookmarks, email links, and sessions without referrer data.
The razor-thin 3-percentage-point gap between direct (48.6%) and search-dependent (45.6%) traffic means that nearly half of Americans looking for government services still rely on a search engine to get there. For government agencies, this makes both memorable URLs and search engine visibility critical for ensuring citizens can find and access services.
The Phone-First Citizen: 43.5% Access Government Services on Mobile
The device breakdown in the federal data tells a clear story about how the relationship between citizens and government is shifting. 43.5% of visitors access federal government websites from a mobile phone, with an additional 1.2% on tablets, bringing the total non-desktop share to 44.7%.
| Device Category | % of Total Users | Rank |
| Desktop | 55.1% | 1 |
| Mobile | 43.5% | 2 |
| Tablet | 1.2% | 3 |
| Smart TV | 0.2% | 4 |
The 11.6-percentage-point gap between desktop (55.1%) and mobile (43.5%) represents both a milestone and a warning. It signals real progress in digital inclusion, as more Americans can reach government services from the device they carry with them. But it also raises a question: are government websites, many of which were designed for desktop use, built to handle complex tasks like tax filing, benefit applications, and healthcare enrollment on a mobile screen?
The data does not answer that question directly, but the scale of mobile access makes it unavoidable. When 43.5% of your users are on a phone, mobile optimization is not a feature request. It is a baseline requirement.
The 0.9% Who Are Still Using Internet Explorer
Tucked into the browser data is a detail that says as much about digital equity as any headline number. According to the federal analytics, 0.9% of visitors to government websites are still using Internet Explorer, a browser that was officially discontinued in June 2022.
| Web Browser | % of Total Users | Rank |
| Chrome | 56.1% | 1 |
| Safari | 28.2% | 2 |
| Edge | 8.3% | 3 |
| Safari (in-app) | 1.9% | 4 |
| Firefox | 1.9% | 5 |
| Android WebView | 1.1% | 6 |
| Samsung Internet | 1.0% | 7 |
| Internet Explorer | 0.9% | 8 |
| Opera | 0.3% | 9 |
| Other | 0.2% | 10 |
| Mozilla Compatible Agent | 0.1% | 11 |
That 0.9% is not a rounding error. It represents a meaningful population of Americans who may be accessing federal services on outdated workplace systems, older personal computers, or institutional machines that have not been updated. These users are likely to include older Americans, individuals in institutional settings, and people using legacy workplace technology, populations that often have the greatest need for government services.
At the other end of the spectrum, Chrome (56.1%) and Safari (28.2%, plus 1.9% in-app) together account for 86.2% of all government website traffic. This browser concentration means that government web teams are effectively building for two rendering engines, with a long tail of legacy and niche browsers still in active use.
The Bigger Picture: Five Numbers That Define Digital Government Access in 2026
The federal government’s own analytics data, publicly available at Analytics.usa.gov, paints a picture of a population in transition. The numbers tell a story of citizens who are increasingly mobile, increasingly reliant on search engines, and beginning to use AI tools to navigate public services.
| Metric | Figure | What It Signals |
| Direct navigation | 48.6% | Nearly half already know where to go |
| Search engine dependence | 45.6% | Nearly half still need help finding services |
| Mobile access | 43.5% | Mobile optimization is a baseline requirement |
| Chrome + Safari dominance | 86.2% | Two browsers define the testing surface |
| ChatGPT referral traffic | 0.7% | AI is now a measurable access channel |
| Internet Explorer usage | 0.9% | Legacy access remains a real equity issue |
None of these numbers exist in isolation. The 43.5% filing on mobile may overlap with the 45.6% who depend on search engines. The 0.9% on Internet Explorer may be among the citizens with the most urgent need for government services. And the 0.7% arriving via ChatGPT represents a trend line, not a destination.
What the data makes clear is that the digital front door to American government is no longer a single door. It is dozens of entry points, across devices, browsers, and now AI assistants, each serving a different slice of the public. The question for government agencies is whether their digital infrastructure is built for the citizens they actually have, not the ones they assumed they would.
Methodology
Data for this analysis was sourced from Analytics.usa.gov, the U.S. federal government’s official digital analytics platform operated by the Digital Analytics Program (DAP). All figures are publicly available and independently verifiable. The analysis examined aggregate traffic patterns across federal government websites, including session source attribution, device category distribution, browser type usage, and referral patterns. The data was collected on January 27, 2026, reflecting the 30-day rolling period from December 28, 2025 through January 26, 2026. This analysis is based solely on publicly available traffic and usage metrics and does not evaluate the quality, security, or effectiveness of any government service or private platform.
Note: “Direct” traffic includes users who typed the URL directly, used bookmarks, clicked links from emails or documents, or accessed sites through channels that don’t pass referrer information. It represents users who reached government sites without an intermediary search or referral page, though the specific navigation method varies.
Data Sources
U.S. Federal Government Official Web Analytics: https://analytics.usa.gov/
Research Dataset: https://docs.google.com/spreadsheets/d/1qL_SB6F_pL168j-x3abESuFBjaSuO24OmydD9KFCnls/edit?gid=1367810881#gid=1367810881
Study by: VWO.com (https://vwo.com/). A leading website optimization platform that helps businesses run A/B tests and improve their digital experiences.
The information provided in this article is for educational and awareness purposes only, and is provided in good faith. We make no representation or warranty of any express or implied kind with regard to the adequacy, validity, completeness, or reliability of the information provided herein. All trademarks, brand names, and/or logos are properties owned by their respective owners.