Attention marketers: New research using 2.7 million data points from the 2026 Winter Olympics reveals how AI systems shape and sustain narratives around brands, athletes, and organizations.
This has huge implications for your brand and many others.
What did researchers from Seer Interactive discover?
When a user asks ChatGPT, Gemini, or Perplexity about a company, they are not simply retrieving current facts. They complete the story that has formed over time.
So, if the dominant narrative around your brand was established months or years ago, through a Glassdoor review, a news story, or an analyst report, for example, AI will likely continue to tell that story even if it changes.
The researchers at Seer Interactive call this “narrative appeal.”
The concept arose from an ambitious study of artificial intelligence search behavior. John LovettVP of Analytics at Seer Interactive, and his team used the 2026 Winter Olympics as a real-time proving ground, tracking six major AI platforms, including ChatGPT, Gemini, Google AI Mode, AI Overviews, Perplexity, and Meta AI, over a nine-week period.
The team asked the same questions across all platforms every day, ultimately generating 2.7 million data points on how AI systems display, cite, and suppress information.
The results are shown in Artificial Intelligence For B2B Marketers Summit, It has huge implications for your brand.
Experiment: When the favorite loses, the AI still says he won
The Olympics provided a rare, controlled environment for Lovett and his team to experiment with AI research: predictable narrative arcs, high-frequency breaking news, global media coverage, and clear, verifiable results.
One pattern continued to emerge: Before the event, a consensus story formed around a favorite athlete or team. When the actual outcome deviated from this expected arc, the researchers asked narratively framed questions, the kind an average user might type into an AI chatbot. The AI systems continued to complete the expected story. With confidence.
“The wording of the question determines whether you get the current fact or a previously completed narrative based on parametric (training) knowledge,” Lovett explained. “If your brand exists within a dominant industry story, AI may continue to tell that story regardless of your latest move.
“This represents a risk and an opportunity at the same time. “You can influence him.”
For brands in rapidly changing industries or brands trying to reposition, this is key to understanding and addressing.
It’s hard to shake off the past in AI
To illustrate how Narrated Gravity can play out in a business context, Lovett points to Seer Interactive’s own experience.
The company discovered that several MBAs were showing a years-old negative post on Glassdoor when users asked about Seer, framing it as evidence of an employee retention problem. The models were not incorrect in reporting the presence of the review. But they were presenting a single data point as a specific fact about the organization.
“If a brand has a good reputation for something, it’s hard to shake,” Lovett said. “The narrative appeal of an MBA is like having your reputation precede you.”
Seer wrote a blog post directly addressing the issue in a deliberate attempt to introduce a counter-narrative into the information ecosystem. However, the broader lesson is structural: AI systems try to provide “balanced” information, and in doing so, they can give disproportionate weight to isolated negative signals simply because those signals exist and have been disseminated.
For marketers, this reframes reputation management in an important way. It’s no longer just a question of what your brand says about itself. This is what the training data says about your brand and whether you’ve created enough authoritative content that counters any of the negative things the AI ”knows.”
The Escher Principle: You can’t buy your way past the moment
The research revealed a second key finding that Lovett called the Escher Principle: Events amplify what already exists. They do not create existence out of nothing.
During the Olympic Games, some athletes made real news with their results, records and star moments echoed across the world’s media. But if these athletes lack a pre-existing digital footprint — owned content, third-party verification, community discussion — AI systems have largely failed to surface them in response to relevant queries. News can’t build a foundation. It could only amplify one that already existed.
The data showed three vision signals that stack up when present together:
- Entity authority (you own your entity definition)
- Third party verification (confirmed by others)
- Community discussion (audience reinforces it)
These signals are sequential or interchangeable.
“The entity authority shuts everything down,” Lovett said. “You own your being first, third parties validate your personality second, and community discussion reinforces you third. Skip the first step and the others won’t double.“
The gap between having one signal versus all three is huge: 7.8x the average AI signals.
It’s a cumulative advantage: AI entity detection with established authority. This insight generates new third-party coverage. This coverage is due to more authority. The system rewards previous attendance, then doubles it.
“The authority you build today is cheaper than the authority you will need tomorrow to reach the same position, because each cycle of AI-mediated searches raises the price of entry for everyone who waited,” Lovett said.
What this means for B2B marketers now
the State of Business AI 2026 Report A study published by SmarterX found that only 3% of respondents cite AI-powered search as a trend they are following closely, suggesting more attention needs to be paid to how AI systems are reshaping discoverability.
Meanwhile, 40% of professionals say agents and agentic AI are their top emerging trends, suggesting that teams are focusing on AI production capabilities while their AI vision may be quietly eroding.
For B2B brands, the risk is particularly acute. Purchasing decisions are increasingly made through AI-generated summaries. A buyer asking AI about sellers in your category will receive a rich answer, and if your brand lacks entity authority, that answer will not include you, no matter how much you spent on paid search last quarter.
Lovett session in Artificial Intelligence for B2B Marketers Summit It will provide attendees with a validated framework for AI visioning, concrete tactics for improving citation, and a clear model of how real-time events can reshape what AI systems “know.”
To learn more about the AI for B2B Marketers Summit or to register, visit: https://www.marketingaiinstitute.com/events/ai-for-b2b-marketers-summit
(function(d, s, id) {
var js, fjs = d.getElementsByTagName(s)(0);
if (d.getElementById(id)) return;
js = d.createElement(s); js.id = id;
js.src = “//connect.facebook.net/en_GB/sdk.js#xfbml=1&version=v3.0”;
fjs.parentNode.insertBefore(js, fjs);
}(document, ‘script’, ‘facebook-jssdk’));
