• by Laurie Sullivan , Staff Writer, 7 hours ago

Organic impressions -- not paid media ad impressions -- have become the currency for brand citations that serve up in artificial intelligence (AI)-based engines.

Recommendations have become the first "impression" a brand makes with a consumer. But can that first impression be measured and attributed in a similar way?

Brand discovery has moved from search engines to AI assistants as consumers rely more heavily on ChatGPT, Gemini and Perplexity for answers on what to purchase and where to buy it.

AI assistants return a few brand names with short descriptions while omitting others. Most of these exchanges between humans and AI assistants end without a click.

This forces advertisers to rethink how an AI system describes and recommends a brand.

It has become important for these AI assistants to mention brands and their products in citations, which act like a recommendation.

Marketers can measure "mention" rates in AI engines through a product developed by Clinch, an AI-powered company. On Wednesday Clinch introduced what the company calls "Generative Engine Optimization (GEO)," which integrates into its "Flight Control" platform and provides measurement that is not done in the same way as a typical paid-ad impression.

GEO provides insights to determine which type of first impression a marketer's brand may make on a consumer based on the creative, and the campaign strategy.

GEO tracks and optimizes how AI engines like ChatGPT, Google Gemini and Microsoft Copilot cite, describe and recommend brands in user queries.

“Generative AI is becoming the narrator at the top of the funnel, shaping how brands are described, compared, and recommended before a consumer ever takes the next step,” said Sam Jones, senior vice president of performance strategy at Canvas Worldwide, adding: “In high-consideration categories like automotive, the narrative can decide who earns real consideration and who gets filtered out early.”

Retail has already begun to experience this shift. AI referral traffic to U.S. retail sites grew 693% year-over-year during the 2025 holiday season, according to Adobe Digital Insights, looking at results from January 2026.

The data shows those AI referrals managed to convert 31% more effectively than non-AI traffic over the same period.

Forty percent of marketers already report brand-visibility growth from AI assistants such as ChatGPT, Gemini, and Perplexity, according to Fractl data from 2026.

"Rather than a brand buying their way to ranking high based on keywords, they are trying to make themself more discoverable by influencing the attention of their brand," Charel MacIntosh, global head of strategic partnerships at Clinch, told MediaPost, adding that the attention comes from the information an AI engine serves up based on the user query. 

The countable metrics for GEO generally fall into visibility, sentiment, comparison, and mention rate. The mention rate is defined as the average percentage of prompts in which a brand appears in AI generated responses. 

Marketers cannot count an impression because an impression is a discrete served event in a log that the advertiser can audit, according to Clinch. An AI recommendation is probabilistic and personalized, so the same question returns different answers to different people at different moments, and there's no fixed number of times a brand is shown. It's measured by sampling, closer to share of voice than to an ad-server count

As these AI platforms add paid placements like OpenAI's ads product, those will be measurable as true impressions in the classic sense, so a paid layer is forming alongside the earned one that GEO measures today.

Marketers can use GEO insights to determine the message, creative, and campaign strategy across channels to create a consistent loop between being discovered in AI engines and the execution of an organic brand strategy without paid media.

Key measurable signals have become the brand's visibility in AI engines based on what large language models (LLMs) learn about the brand.

The key is how the brand shows up, understands the citations and sources that drives consumer perception of the brand -- and how it applies those insights directly to the decisions that make up content and creative for campaigns. It is important for the brand’s message to remain consistent. 

When the AI engine surfaces citations, sources and topics, this determines how models understand and describe the brand. Inaccurate descriptions of brands happens quite often, but that is different topic and discussion.

Brands can then act on those insights served by the AI engine to generate content that is designed to influence how consumers perceive the brand.

The citations also work as recommendations. The same insights can be applied directly to the brand's creative strategy and campaign created inside the Flight Control platform, allowing marketers to align AI-driven consumer perception with the brand's creative messages that are built into the platform. 

Early adopters that had not been identified for these results experienced success with these strategies. For example, a large retailer brand was able to increase its mention rate from 50% to 68% within two months by combining optimized owned-site content with strategic third-party content placement.

The methodology has since been applied across B2B, B2C, and e-commerce customers.