# Gartner has confirmed the battle for authority is happening right now. Is your agency ready for the aftermath? > Source: > Published: 2026-08-14 19:40:00+00:00 # Gartner has confirmed the battle for authority is happening right now. Is your agency ready for the aftermath? ## COMMENTARY | If agencies don't act, the consequences won't be abstract. For decades, government agencies have measured digital success by a familiar yardstick: rank on Google, drive clicks to the .gov site and let citizens navigate from there. That yardstick is breaking. Citizens are no longer starting their search for information with a list of blue links; they're starting it with a single AI-generated answer, and in a growing number of cases, they never click through to a source at all. Gartner has now put numbers behind what many of us have felt building for the past two years. The firm predicts that by 2028, an organization’s organic search traffic will fall by 50% or more as individuals embrace generative AI-powered search, and separately forecasts a 25% decline in traditional search engine volume by the end of this year as AI chatbots and virtual agents take over as answer engines. That's not just a trend. For government, it's a governance problem. **The authority vacuum** When a citizen asks ChatGPT, Gemini or Perplexity how to replace a lost SNAP card or appeal a benefits denial, the AI model doesn't necessarily reach for the agency's website first. It reaches for whatever content it has learned to trust as authoritative. If the agency hasn't structured its content to be clearly sourced, current and machine-readable, the model may pull from a third-party aggregator monetizing confusion or a source from the wrong jurisdiction entirely. This is the "battle for authority." It isn't a battle for rankings anymore. It's a battle over who gets to be the voice an AI system trusts enough to cite. Right now, most government agencies aren't even aware they're in it. Recent research on AI-augmented search behavior found that roughly 60% of internet searches now end without a click, because users are satisfied with the AI-generated summary and feel no need to visit the source. For a private brand, that's a conversion problem. For a government agency, it's a legitimacy problem — because if an AI system is quietly becoming the primary interface between citizens and their government, and the agency has no visibility into what that system is telling people, the agency has effectively lost control of its own message. Three forces are compounding the problem for public sector organizations specifically: **The information ecosystem citizens rely on is already under strain.** Gartner's broader research on citizen trust found that a majority of people believe the quality of online information sources has degraded in recent years, driven by misinformation, low-quality content and bot-generated noise. Citizens are being pushed toward AI assistants as a shortcut past that noise — which only raises the stakes for making sure those assistants are pulling from the right government source when they answer.**AI models reward structure, not just accuracy.** Being factually correct is no longer sufficient. Generative engines favor content that is written in plain, question-and-answer language, organized with clear headings and short paragraphs, dated and refreshed regularly and backed by verifiable citations and consistent cross-references. A perfectly accurate policy page buried in a dense PDF with no update date is far less likely to be surfaced than a clearly structured FAQ — even if the FAQ is hosted by a third party with a looser relationship to the truth.**Small visibility gaps compound into large trust gaps.** Emerging data suggests that even modest improvements in AI citation share can translate into disproportionate downstream effects — a meaningful increase in how often an agency is cited by generative engines has been associated with a substantially larger lift in actual service enrollment. The inverse is just as true: agencies that lose citation share don't just lose traffic, but they also lose the ability to correct misinformation before it reaches citizens who need accurate, time-sensitive guidance on benefits, health alerts or public safety. **What the aftermath looks like** If agencies don't act, the consequences won't be abstract. They'll show up as citizens receiving incorrect eligibility guidance synthesized from outdated or out-of-jurisdiction sources; a widening trust gap as AI-cited misinformation goes unanswered because no one at the agency is monitoring what AI platforms are actually saying; wasted resources as digital teams keep optimizing for a search paradigm that's shrinking, while ignoring the one that's growing; and, over time, a slow transfer of "official voice" status from the agency itself to whichever aggregator, advocacy group or commercial site the AI models learn to trust instead. None of this is hypothetical anxiety about the future. It's a description of where the traffic and trust data already point. **Building authority before the vacuum fills itself** The good news is that authority in the GEO era is buildable. It's an evidence architecture, not a mystery. Agencies that want to remain the trusted source need to treat AI visibility as seriously as they've historically treated public affairs and constituent communications, which means monitoring what AI platforms are saying about their programs and policies, not just what's ranking on Google; restructuring core content — FAQs, benefit explainers and safety notices — into the answer-first, clearly dated, plainly written format that generative engines are built to extract and cite; and closing the loop between visibility and outcomes, connecting AI citation data to actual service utilization so digital teams can prioritize the highest-impact pages first. This is precisely the gap we are trying to close for government digital teams — giving agencies visibility into how they're represented across AI platforms, insight into how AI bots are actually using their content and prescriptive, actionable recommendations for closing the authority gap before a citizen encounters bad information instead of good. **The mandate is already here** Mandates like the 21st Century IDEA and a growing body of state and local digital service regulations already require agencies to deliver clear, accessible information to the public. Those requirements didn't anticipate a world where the primary access point to that information is a generative AI model rather than a webpage — but the underlying obligation hasn't changed. Accessibility and accuracy now have to be engineered for machines that summarize on the public's behalf, not just for the humans who used to click through. The battle for authority Gartner is describing isn't coming; it's underway. Agencies that recognize this now — and start building the structured, monitored, machine-readable authority that AI platforms are built to trust — will be the ones citizens can still find, and still believe, when they ask an AI assistant instead of a search engine. The ones that wait will find out the hard way what happens when someone else becomes the voice of their agency instead. *Brian Chidester is the Head of Industry Strategy for the Public Sector at Adobe.*