BrightEdge AI Catalyst analysis reveals that ChatGPT, Perplexity, Gemini, Google AI Mode, and Google AI Overviews operate with fundamentally different editorial personalities when selecting the sources they cite. At the same time, the brands named in AI-generated answers remain far more consistent across engines than the sources those engines use to construct those answers. The gap between how engines source and what engines recommend is the single most important pattern for any brand building an AI search strategy.

The prevailing assumption in AI search is that each engine requires its own playbook because each engine behaves differently. The data confirms the engines do behave differently, in some cases by close to two orders of magnitude. But the consistency on the output side, which brands get named in the final answer, tells a different story. The playbook does not need to be fragmented by engine. It needs to be organized by source layer.

This is the latest installment in our BrightEdge AI Catalyst research series. We analyzed citations and brand mentions across ChatGPT, Perplexity, Gemini, Google AI Mode, and Google AI Overviews, drawn from prompts spanning ten industries including B2B technology, education, entertainment, finance, healthcare, insurance, restaurants, travel, and ecommerce. The patterns that emerged are directly relevant to any brand planning AI visibility at scale.

Data Collected

Key Finding

AI search engines are often discussed as if they behave similarly because they produce a similar kind of output: a synthesized answer with citations. The BrightEdge AI Catalyst data shows that behind the surface, the five engines pull from meaningfully different parts of the web. The share of citations coming from authoritative sources ranges from 10% to 26%, depending on the engine. The share coming from user-generated content ranges from 0.2% to 18%, roughly a 90x spread across engines answering the same categories of questions. Despite that divergence in sourcing, the brands those engines recommend cluster in a much tighter range. Pairwise top-100 overlap in named brands across engines falls between 36% and 55%, a 19-point spread, while pairwise top-100 overlap in cited sources ranges from 16% to 59%, a 43-point spread. Source agreement between any two engines varies widely and inconsistently. Brand agreement is consistently steady. The implication for brand strategy is that the path AI takes to reach its answer matters less than most strategies assume, but being present across the three distinct source layers that feed those paths matters more than strategies typically account for.

Five AI Engines, Five Sourcing Personalities

Gemini functions as a formal institutional recommender. Gemini shows the strongest bias toward authoritative sources of any engine in the dataset. Approximately 26% of Gemini's citations come from government domains, academic institutions, and major industry institutional bodies combined. UGC and social content makes up only 0.2%. The authority-to-UGC ratio is roughly 130 to 1, the highest in the study. Gemini also shows the highest .gov share of any engine at roughly 13%, paired with a .org share of 23%. The engine behaves as a conservative, list-oriented recommender that leans on trusted institutional voices and tends to produce longer, more inclusive brand lists than other engines.

ChatGPT acts as a long-tail editorial engine. ChatGPT cites the flattest source distribution of any engine in the study. Its top 10 most-cited domains account for only 18.5% of total citations, meaningfully lower than Perplexity (26.7%), Gemini (26.3%), or AI Mode (19.4%). ChatGPT also has almost no UGC presence (0.5%) and pulls heavily from government and .org domains (12% and 20% respectively). The engine reads as a formal editorial assistant with a long, diverse tail of corporate, institutional, and government sources.

Perplexity behaves like a research librarian. Perplexity concentrates more of its citations in institutional medical, government, encyclopedic, and medical publisher sources than any other engine. Combined, those four categories account for approximately 30% of Perplexity's citations. Perplexity shows the highest share of .edu citations (3.2%) and the highest share of international country-code domains (4.4%) in the dataset, reflecting a more formal and globally sourced material mix. It also names brands earliest of any engine, with 86% of its brand mentions landing in position 5 or earlier. Perplexity behaves like an engine that commits to a short, authoritative shortlist rather than producing an exhaustive list.

Google AI Mode operates as a broad commercial aggregator. Google AI Mode pulls from a wider catalog of unique domains than most other engines, with a long-tail distribution that spreads citations across far more sources than its siblings. It also distributes its citations more evenly across source types than any other engine in the study, showing the strongest mix of review aggregators, finance data sources, and news media citations in the dataset. UGC exposure is moderate at roughly 7%, well above ChatGPT or Gemini but well below AI Overviews. AI Mode's top 10 citation concentration is among the lowest at 19.4%, reinforcing its identity as a long-tail, balanced commercial surface.

Google AI Overviews is a UGC-first engine. Google AI Overviews stands apart from every other engine in the study. Approximately 17.5% of its citations come from user-generated content platforms, 35x higher than ChatGPT (0.5%) and 87x higher than Gemini (0.2%). A single video platform accounts for roughly 10.6% of all AI Overviews citations on its own, and a single forum platform adds another 2.9%. Authoritative sources, including government, academic, and major institutional bodies, account for only 9.5% of AIO citations combined. AI Overviews is the only engine in the dataset where UGC citations outweigh authoritative citations.

Authority Share Versus UGC Share, by Engine

The Two Google Engines Are Not the Same Engine

Among the five engines studied, the two most similar are Google AI Mode and Google AI Overviews, with a top-100 citation overlap of roughly 59%. But Gemini, also a Google product, behaves very differently from its siblings. Gemini's top-100 citation overlap with AI Mode is only 27%, and with AI Overviews only 34%. Gemini actually has more in common with ChatGPT (39% overlap) than with the Google search-embedded surfaces. In practical terms, "Google AI" is not one thing. The search-embedded surfaces lean heavily on commercial and UGC content, while standalone Gemini behaves like a conservative, authority-heavy reference engine. Any brand strategy that treats all three Google AI surfaces as interchangeable will miss the actual sourcing patterns driving visibility on each.

The Brand Convergence Signal

The most consequential finding in the study is not the divergence in sources. It is the convergence in brand recommendations despite that divergence. Pairwise top-100 overlap in cited sources across engines ranges from 16% to 59%, a 43-point spread. Pairwise top-100 overlap in named brands ranges from 36% to 55%, a 19-point spread. In every pairwise comparison, brand overlap falls in a tighter, more predictable range than source overlap. The engines disagree substantially and inconsistently about where to pull information from. They agree more consistently about which brands belong in the final answer. That pattern is what makes a unified strategy viable across all five engines, rather than five separate playbooks.

Sentiment Is Overwhelmingly Positive Across Every Engine

Brand sentiment in AI-generated answers skews positive on all five engines, but not uniformly. Gemini is the most positive at roughly 96% positive sentiment, with only 0.3% negative. ChatGPT sits at 94% positive with effectively zero negative mentions. Perplexity shows the highest neutral share at 11%, consistent with its more journalistic, reference-oriented posture. The Google search-embedded surfaces (AI Mode at 93% and AI Overviews at 89%) show slightly higher negative sentiment (1.7% and 2.1%), which reflects their deeper pull from UGC and commercial commentary sources where critical framing more commonly appears. Across the dataset, negative brand mentions remain a marginal share of total volume, which reinforces that visibility in AI answers is almost always presented in a positive or neutral frame.

What Marketers Need to Know

AI engines pull from three distinct source layers, and every engine uses all three. Authoritative sources include government, academic, and major industry institutional content. Commercial and editorial sources include review sites, comparison content, trade press, news media, finance data, and retailer listings. UGC includes video content, forum threads, community discussion, and creator coverage. No engine uses only one layer. The engines differ in how they weight each layer, not in whether they use it. A brand visibility strategy built around only one layer, no matter which layer, will underperform on engines weighted toward the other two.

Authority is category-relative. "Authoritative" does not mean .gov or .edu for every brand. Not every company can or should aim to be cited by federal agencies or academic institutions. Every category has its own authoritative layer: trade associations, analyst firms, expert publishers, standards bodies, professional associations, and institutional voices trusted within the vertical. The strategic question is which authoritative sources serve as the backbone of AI citations in your specific category, and whether your brand is covered by those sources.

Commercial and editorial presence is the widest visibility lever. Across all five engines, the brand/corporate and commercial/editorial source layer accounts for the largest share of citations, ranging from roughly 37% on Gemini to 51% on AI Overviews. Review sites, comparison content, trade press, retailer listings, and finance data are the sources AI most frequently reaches for. Investment in PR, trade coverage, review site visibility, and category comparison content translates into visibility across every engine, not just one.

UGC is non-negotiable for AI Overviews and still meaningful elsewhere. The AI Overviews surface draws roughly 18% of its citations from user-generated content, but UGC is not zero on other engines either. Perplexity pulls 1.5% of its citations from UGC, AI Mode pulls 7%, and both represent real retrievable impressions in categories where community and creator content is strong. A UGC strategy does not mean "produce short-form video." It means understanding which videos, forum threads, and community discussions AI is already citing in your category, and being present in that conversation with authority.

Weight investment based on which engines matter most to your buyers. The three-layer framework is universal. The emphasis is not. A B2B SaaS brand whose buyers rely heavily on ChatGPT and Perplexity will prioritize authority and commercial coverage, with UGC as a supplemental layer. A consumer brand whose buyers use AI Overviews heavily will prioritize UGC and commercial presence, with authority as reinforcement. Brand tracking at the engine level, not just in aggregate, is how those priorities get set and validated.

Engine overlap patterns should inform where you measure first. The two Google search-embedded surfaces share roughly 59% of their top-cited sources, so visibility gains on one frequently translate to the other. Gemini behaves more like ChatGPT than like its Google siblings, so brand teams should not assume that a Gemini strategy is a Google strategy. These overlap patterns are not intuitive, and brands that map their measurement plan against actual engine behavior will catch gaps that aggregate tracking hides.

Technical Methodology

Key Takeaways

BrightEdge AI Hyper Cube analysis of auto prompts across Google AI Overviews and ChatGPT reveals that non-branded queries dominate the top of the purchase funnel -- and that AI recommends brands in nearly every response, whether shoppers ask for one or not.

Every day, car shoppers turn to AI with questions about vehicles, financing, reliability, and deals. But a closer look at how those prompts are structured reveals something that challenges a core assumption about AI search behavior in automotive: a substantial share of auto-related AI prompts contain no brand name at all. And yet brands are being recommended in the AI-generated answers almost every single time.

We used BrightEdge AI Hyper Cube to map the full AI prompt universe across the top auto brands in both Google AI Overviews and ChatGPT. Prompts were divided into three stages of the purchase funnel -- informational, consideration, and transactional -- and analyzed for whether they contained a brand name. We then examined whether brands appeared in the AI-generated answer regardless of whether the prompt named one.

Data Collected

Key Finding

Automotive search strategy has long been organized around branded intent. The assumption is that consumers who are ready to act know which brand they want, and that non-branded queries belong to the awareness stage where brand influence is limited. The prompt data challenges both assumptions. Non-branded prompts represent a significant share of auto AI search volume at every funnel stage -- and AI is actively recommending brands in response to those prompts 97% of the time. The implication is direct: auto brands are not just competing for visibility when shoppers search their name. They are competing to be the brand AI recommends when shoppers do not search for anyone.

Four Patterns Across the Auto AI Purchase Funnel

At the informational stage, branded and non-branded prompts are nearly equal in volume. On Google AI Overviews, non-branded prompts account for 48% of informational auto prompt volume. Shoppers at this stage are asking about car maintenance, towing capacity, charging infrastructure, fuel economy, and vehicle comparisons without naming a specific brand. These are not low-intent queries. They are the first moment of AI-assisted discovery, and brands are being named in AI answers to these prompts 97% of the time. The brand that earns placement in informational AI answers is setting the consideration set before the shopper has explicitly formed one.

Brand intent increases measurably as shoppers move toward purchase. By the consideration stage, branded prompt volume climbs from 52% to 64% on Google AI Overviews -- a 12-point shift that reflects shoppers narrowing their options and beginning to research specific makes, models, trim levels, and lease deals. The non-branded share at consideration still represents more than one-third of prompt volume. Prompts like "most reliable car brands," "best used cars to buy," and "luxury SUV brands" contain no brand name but generate AI responses that name multiple brands, rank them, and editorially favor some over others. For brands not appearing in those answers, consideration-stage visibility is effectively zero.

The transactional stage shows the clearest brand concentration on Google AI Overviews, where 67% of prompt volume is branded. Shoppers at this stage are pricing specific models, searching for dealer locations, comparing lease offers, and requesting test drives. They have done their consideration work. But one-third of transactional prompt volume on Google AI Overviews still contains no brand -- prompts like "car dealership near me," "0% finance car deals," and "test drive" -- and brands are still being recommended in the AI-generated responses to those queries.

On ChatGPT, the transactional stage behaves differently and in a way that is strategically significant. Despite transactional prompts being the stage most associated with brand-specific intent, 70% of transactional auto prompt volume on ChatGPT is non-branded. Prompts like "used cars for sale," "work trucks for sale," and "what car has the best rebates right now" are purchase-intent queries that name no brand. ChatGPT is generating brand recommendations in response to all of them. This pattern suggests that ChatGPT users at the transactional stage are more likely to be delegating the brand decision to AI rather than arriving with a brand already selected.

Branded vs. Non-Branded Prompt Volume by Funnel Stage

The 97% Signal

Across all funnel stages and both platforms, 97% of non-branded auto prompts resulted in auto brands being named in the AI-generated answer. This finding reframes where the competitive battle in AI search actually takes place. Whether a shopper types a brand name or not, AI is selecting brands and presenting them with varying degrees of prominence and sentiment. The prompt is not the battleground. The answer is. Brands that are not present in AI-generated responses to non-branded prompts are absent from a substantial portion of the consideration and purchase journey -- even though no shopper explicitly excluded them.

What Marketers Need to Know

Non-branded prompt volume is not awareness-stage noise. Nearly half of all informational auto AI prompt volume contains no brand name, and those prompts are generating brand recommendations at a 97% rate. A visibility strategy built only around branded query performance is measuring the wrong thing.

ChatGPT transactional behavior in auto is fundamentally different from Google AI Overviews. The 70% non-branded transactional volume on ChatGPT suggests a platform where shoppers are more likely to ask AI to help them decide rather than arriving with a brand already chosen. Content and product pages that can be surfaced in response to generic purchase-intent queries need to be AI-accessible on this platform.

AI is building consideration sets before shoppers do. The brands that appear in AI answers to informational non-branded queries are establishing familiarity and preference before a shopper has consciously begun comparing options. Informational content -- reliability data, comparison content, ownership cost breakdowns -- needs to be optimized for AI citation, not just organic ranking.

Prompt share does not equal answer share. A brand can be named in a prompt without being recommended prominently in the answer, and a brand can be absent from prompts entirely while appearing consistently in AI-generated responses. Understanding where your brand appears in AI answers -- across branded and non-branded prompts at every funnel stage -- is a distinct and necessary measurement capability.

Technical Methodology

Key Takeaways

BrightEdge AI Hypercube analysis reveals that healthcare-related ChatGPT prompts are dominated by transactional intent, with patients and members using AI not just to research conditions but to find care, price procedures, self-treat, and act on their benefits.

People turn to ChatGPT with health questions every day. But a closer look at the prompts reveals something that challenges a core assumption about how AI search works in healthcare: the majority of transactional prompts aren't about learning. They're about acting. People aren't just researching conditions. They're finding providers, pricing care, managing symptoms, and trying to unlock their benefits.

This is the second installment in our AI Hypercube transactional intent series. Last week we analyzed finance. This week we turned the same methodology on healthcare, looking at queries tied to top U.S. health brands and filtering for transactional intent. The patterns that emerged are consistent, significant, and directly relevant to any healthcare brand thinking about AI visibility strategy.

Data Collected

Key Finding

Healthcare is widely treated as a research-heavy, high-trust informational vertical in search strategy. The assumption is that people use AI to learn about conditions, treatments, and providers before taking action elsewhere. The prompt data tells a different story. Transactional intent is present throughout ChatGPT healthcare queries, from finding a family doctor to pricing a dental procedure, checking GLP-1 coverage, self-treating symptoms, and figuring out how to spend an OTC benefits card. Unlike finance, where transactional intent clustered around a single action (applying), healthcare transactional intent fragments across the entire patient and member journey. The implication for healthcare brands is direct: an AI visibility strategy built only around condition content and symptom pages is incomplete.

Five Transactional Intent Clusters in ChatGPT Healthcare Prompts

Finding care is the single largest cluster of transactional healthcare prompts in ChatGPT, accounting for roughly 55% of identified transactional volume. A single prompt, "family doctor near me," drives over 7,000 monthly queries on its own. Variations include urgent care searches, specialist lookups, and "accepting new patients" queries across dental, primary care, and specialty medicine. These are not research prompts. They are front-door prompts from people ready to book. The shift is significant: AI is becoming the discovery layer for care access, replacing traditional provider directory searches and health plan "find a doctor" tools.

Cost shopping accounts for approximately 17% of transactional healthcare volume. People are using ChatGPT to price procedures before committing to them. "How much do veneers cost" drives over 1,600 monthly prompts. Queries for the cost of tooth extraction, dental implants, eye exams, and contact lens fittings appear repeatedly. Dental and vision dominate this cluster, which is consistent with how consumers experience healthcare economics: these are the procedures they pay for most directly. Cost-shopping prompts signal someone who has already decided to act and is now comparing where to do it.

Self-treatment and symptom action represents approximately 16% of transactional volume and is the most distinctive healthcare pattern in the dataset. Prompts like "how to cure neck pain fast," "how to relieve period cramps fast," "how to fix gingivitis," and "how to get rid of cold sores fast" reflect people asking ChatGPT for an action to take, often in place of seeing a provider. The AI is functioning as a first-line care substitute. This is a pattern that has no direct analog in finance, and it reshapes what "content strategy" needs to mean for health systems, pharma brands, and retail health players.

Insurance coverage and enrollment prompts account for approximately 9% of transactional volume, with the pattern heavily concentrated around GLP-1s and weight loss medications. Queries like "Is Ozempic covered by Medicare," "Does Aetna cover weight loss medication," "Will Aetna cover Zepbound," and "How to get insurance to cover GLP-1" reflect users trying to unlock access to a specific drug through their benefits. Parallel queries about buying health insurance on the open market, Medicaid enrollment, and plan selection extend the same access-seeking pattern into the coverage itself.

Benefits card utilization accounts for approximately 2% of volume but represents the most maximally transactional behavior in the dataset. Prompts like "Where can I use my Humana spending account card," "Can I buy toilet paper with my OTC card," and "Can I use my OTC card on Amazon" are from members mid-transaction, trying to determine in real time whether a specific purchase is covered by their benefits card. These are not questions about healthcare. They are questions about completing a purchase. The intent state is closer to a checkout decision than a search query.

Emerging Signal: Agentic Scheduling and Plan Shopping

As in finance, a small but directionally significant pattern of agentic prompts appeared in the healthcare data. These are prompts written not as questions but as direct instructions: "Find a primary care physician accepting new patients in [city]," "How do I find a Medicare Advantage plan with wellness programs in Maryland and Virginia," "Schedule a pediatric specialist visit in [city]." These prompts do not have traditional keyword search equivalents. They reflect a user treating ChatGPT as an agent capable of initiating care access, not just describing how to find it. The behavior is early, but the healthcare version of agentic intent points at two natural use cases: appointment booking and plan shopping. Both are high-stakes, multi-variable decisions that consumers have historically handled through call centers or paid agents, and both are exactly the kind of task an AI agent is well-suited to handle.

Intent Cluster Distribution

The Access Signal

The distinctive finding in healthcare is that transactional intent is distributed across the entire journey of gaining access: access to a provider, access to a procedure at a price the consumer can afford, access to a drug through coverage, access to a benefit the member has already paid for. In finance, intent concentrated around a single action. In healthcare, it fragments across five distinct access moments, each with its own content, page type, and operational owner inside a health organization. That has direct implications for AI visibility strategy. Owning educational content on a condition is no longer sufficient. Brands need to ensure AI can reach the parts of their ecosystem where access decisions get executed: provider directories, scheduling systems, cost and procedure pages, formulary and coverage pages, and member resources.

What Marketers Need to Know

Transactional intent in ChatGPT is real and present in your healthcare category right now. The assumption that AI search is a top-of-funnel channel for healthcare does not hold. People are finding providers, pricing procedures, checking coverage, and acting on benefits inside ChatGPT. Content strategy and AI visibility strategy need to account for where people are in the decision process, not just what condition they are researching.

AI agents need access to the parts of your site where action happens. When someone asks ChatGPT to find a doctor, price a procedure, or figure out what their benefits card covers, the AI needs to be able to surface your provider directory, scheduling pages, cost pages, formulary, and member resources. If your AI-accessible content footprint consists primarily of condition pages and symptom explainers, you are optimizing for the wrong moment.

Coverage and access content drives decisions at the moment of truth. The concentration of GLP-1 coverage prompts, OTC card utilization prompts, and in-network provider prompts shows where benefit-gated decisions are being made. These pages have historically been treated as member-portal content, often behind login walls or buried in PDF formularies. That content needs to be citable by AI.

Self-treatment behavior is a signal, not just a threat. People bypassing the care journey to ask ChatGPT for a remedy are showing exactly where your education-to-action content is weakest. Brands that treat self-treatment prompts as a content opportunity, connecting symptom action content to the appropriate next step (telehealth, urgent care, pharmacy, provider booking), can reclaim that moment.

Agentic prompts are an early signal of where this goes. Prompts framed as instructions to book appointments or compare plans do not carry traditional search volume, but they represent the leading edge of AI-facilitated healthcare decision-making. The brands that build AI-accessible scheduling and plan-comparison infrastructure now will own that conversation as it scales.

Technical Methodology

Key Takeaways

People turn to ChatGPT with financial questions every day. But a closer look at the prompts reveals something that challenges a core assumption about how AI search works: a significant portion of finance-related ChatGPT prompts aren't informational at all. They reflect transactional intent. People aren't just learning about financial products. They're trying to act on them.

We used BrightEdge AI Hypercube to analyze finance-related prompts in ChatGPT, focusing on queries tied to the top U.S. banks and financial institutions. The patterns that emerged are consistent, significant, and directly relevant to any financial services brand thinking about AI visibility strategy.

Data Collected

Key Finding

Finance is widely treated as an informational vertical in search strategy. The assumption is that people use AI to research, compare, and learn before taking action elsewhere. But the prompt data tells a different story. Transactional intent is present throughout ChatGPT finance queries, from credit card applications and pre-approval requests to account opening, credit access management, and full agentic requests to initiate financial processes. The implication for financial services brands is direct: an AI visibility strategy built only around educational and informational content is incomplete.

Five Transactional Intent Clusters in ChatGPT Finance Prompts

Credit card applications and shopping represent the single largest cluster of transactional finance prompts in ChatGPT, accounting for roughly 72% of identified transactional volume. Nearly 90,000 monthly prompts reflect people actively comparing or applying for credit cards. The majority are branded, tied to specific co-branded retail cards, airline loyalty products, and hotel programs. A single co-branded retail card drives approximately 66,000 monthly prompts on its own. These are not research prompts. They are decision prompts from people who have already identified what they want and are using ChatGPT to act on it.

Credit card pre-approval is a distinct sub-pattern worth separating out. We tracked 4,465 monthly prompts for pre-approval queries. Pre-approval is the first step in an application process. Someone asking ChatGPT about pre-approval is not exploring options. They are starting the process of applying.

Banking operations prompts reveal a different kind of transactional behavior. Prompts for opening accounts, setting up direct deposit, initiating bill pay, and processing transfers collectively represent approximately 18% of transactional finance volume. These are people using ChatGPT as a step-by-step guide for completing a banking action. The AI is functioning as a procedural concierge, not an information source.

Credit access management prompts, including queries about checking credit scores, unfreezing credit, and resolving credit report issues, account for approximately 4% of transactional volume. The significance here is what these prompts signal about intent state. Unfreezing a credit file is a prerequisite to submitting an application. A person asking ChatGPT how to unfreeze their credit is not casually curious. They are preparing to apply for something.

Agentic prompts represent an emerging pattern that sits apart from the volume-driven clusters. These are prompts written not as questions but as direct instructions: "Help me start a mortgage refinance application," "Find lenders offering competitive rates for a $250,000 loan," "Get me a personalized refinance payment estimate." These prompts do not have traditional keyword search equivalents. They reflect a user treating ChatGPT as an agent capable of initiating or facilitating a financial process, not just describing one. The behavior is early but directionally significant.

Intent Cluster Distribution

The Branded Signal

Branded prompts, those naming a specific financial institution, retail partner, or co-branded product, drive higher average monthly volume than non-branded equivalents. This is counterintuitive. Broad generic queries might be expected to dominate volume. Instead, the data shows people arriving in ChatGPT with a specific brand already in mind, using the AI to help them take action on a decision already made. Brand presence in AI-generated responses matters most at the exact moment of highest intent.

What Marketers Need to Know

Transactional intent in ChatGPT is real and present in your category right now. The assumption that AI search is a top-of-funnel channel does not hold in finance. People are comparing, pre-qualifying, and initiating applications inside ChatGPT. Content strategy and AI visibility strategy need to account for where people are in the decision process, not just what they want to learn.

AI agents need access to the parts of your site where action happens. When someone asks ChatGPT to find a credit card or start a loan application, the AI needs to be able to surface product pages, application entry points, and offer pages. If your AI-accessible content footprint consists primarily of blog posts and educational resources, you are optimizing for the wrong moment.

Branded intent is concentrated at the decision stage. People arriving in ChatGPT with a specific brand in mind are not at the beginning of their journey. They have already done their consideration work. If your brand is not visible at that moment, you are losing customers who were already sold.

Agentic prompts are an early signal of where this goes. The queries without traditional keyword volume are not low priority. They represent the leading edge of how consumers will interact with AI in high-consideration financial decisions. The brands that understand and optimize for this behavior now will own that conversation as it scales.

Technical Methodology

Key Takeaways

Both Google AI Overviews and ChatGPT cite Wikipedia across a wide range of queries. Both engines clearly trust it. But trust is only part of the story. When you look at what Wikipedia is cited alongside in each engine, and the queries where it ranks first organically but still doesn't appear in the AI layer, a more nuanced picture emerges about how each platform actually defines authority and when it chooses to use it.

We used BrightEdge AI Hypercube and DataCubeX to analyze the prompts where Wikipedia is cited across both Google AI Overviews and ChatGPT, the sources that appear alongside it in each engine's responses, and the organic ranking data for tens of thousands of keywords where Wikipedia ranks and an AI Overview is present. The patterns are consistent and instructive.

Data Collected

Key Finding

Google AI Overviews and ChatGPT cite Wikipedia in fundamentally different contexts. In AIO, Wikipedia sits alongside social platforms and community sources. In ChatGPT, it sits alongside institutional authorities and credentialed reference sources. Same citation. Two completely different signals about what each engine thinks authoritative means. And separately, even holding the #1 organic position isn't sufficient for AIO inclusion. The format of the content has to match what the query actually needs.

Same Source. Completely Different Neighborhoods.

When Wikipedia is cited in a response, it doesn't appear alone. Both engines surface multiple sources per response, and the pattern of what appears alongside Wikipedia is strikingly different between the two platforms.

In Google AI Overviews, the most common sources cited in the same response as Wikipedia are YouTube, Reddit, and Quora, appearing in 13%, 9%, and 6% of Wikipedia-cited AIO responses respectively. The broader co-citation set includes news outlets, entertainment indexes, sports media, and community discussion platforms. Wikipedia in this context is functioning as a credibility anchor in a broad, socially-validated ecosystem.

In ChatGPT, the landscape is almost entirely different. Encyclopedia Britannica appears in the same response as Wikipedia in 43% of ChatGPT responses. Merriam-Webster appears in 13%. The remainder of the top co-citations are health publishers, legal reference institutions, and scientific databases.

AIO puts Wikipedia in the company of platforms where people engage with content. ChatGPT puts it in the company of sources people use to verify it. Same citation. Two completely different competitive sets.

Top Co-Citations Alongside Wikipedia - Google AI Overviews

Top Co-Citations Alongside Wikipedia - ChatGPT

Authority and Citation Are Two Different Decisions.

Across tens of thousands of keywords where Wikipedia holds an organic ranking and an AI Overview is present, Wikipedia makes it into the AIO on fewer than half of those queries. That gap is worth examining closely, because the exclusions aren't random.

When Wikipedia is cited in AIO, 75% of the time it holds a top-3 organic ranking. Median organic position: 2. When Wikipedia is not cited, roughly a third of those cases still have Wikipedia sitting at position #1 organically.

The exclusion pattern reveals why. For live sports queries, real-time events, and navigational searches, Wikipedia holds pages on those topics, but AIO needs a live data feed, not a reference article. The content format doesn't fit the query's immediate need, regardless of how authoritative the domain is. A similar logic applies to certain sensitivity-adjacent topics and queries with strong navigational intent.

Ranking reflects topical authority. AIO citation reflects whether the content format can directly serve what the query needs right now. For Google, those are two separate decisions.

What Marketers Need to Know

The competitive set depends on which engine you're in. If you're competing on queries where Wikipedia appears in AIO, you're competing alongside social platforms, community content, and entertainment sources. If you're competing on queries where Wikipedia appears in ChatGPT, you're competing alongside institutional reference authorities. These require different content investments.

Ranking reflects authority. AIO citation reflects usefulness. Google can rank Wikipedia #1 and still not include it in the AIO, because ranking rewards topical credibility while AIO asks a different question: can this content directly answer what the user needs right now? For real-time and navigational queries, an encyclopedia entry can't, regardless of how authoritative the domain is.

The co-citation data tells you who else is in the room. For any query set where Wikipedia shows up, the co-citation patterns give you an accurate picture of the competitive landscape inside that AI response. That competitive set looks fundamentally different in AIO versus ChatGPT, and mapping it for your own category is the starting point for a differentiated citation strategy across both engines.

Content format is a citation variable. AI Overviews make active judgments about whether a piece of content is the right format to answer a specific query type. Authoritative content that isn't structured to serve the query's immediate need may rank highly and still be excluded from the AI layer.

Technical Methodology

Key Takeaways

When someone's ready to buy, these two platforms take very different paths to an answer.

AIO operates inside a commerce-ready SERP. ChatGPT is the whole page. That architectural difference shapes everything about how each platform handles purchase-intent queries and which brands get cited.

When a consumer types a purchase-intent query into Google or ChatGPT, both platforms are trying to do the same thing: give a useful answer. But the path each takes looks remarkably different. And the difference isn't really about the AI. It's about what's around it.

Google AI Overviews sit on top of a SERP that already has Shopping carousels, merchant listings, and organic results. The AI doesn't need to close the transaction on its own. ChatGPT is the whole page. No carousel. No product listing unit. No organic fallback. When someone asks ChatGPT something with purchase intent, the AI has to do all the work itself, including the evaluative work that a full SERP would otherwise distribute across multiple surfaces.

We used BrightEdge AI Hyper Cube to analyse tens of thousands of prompts where the top U.S. retailers appear, tracking mentions, citations, and brand sentiment across both Google AI Overviews and ChatGPT. We filtered to transactional intent to understand how each platform behaves when someone is ready to buy.

The behaviour gap is real. And the architecture explains almost all of it.

Data Collected

Key Finding

Google AI Overviews and ChatGPT handle retail and purchase-intent queries in fundamentally different ways. Not because they have different goals, but because the environments they operate in are fundamentally different. AIO can lean on the SERP's existing commerce infrastructure to do the transactional heavy lifting. ChatGPT cannot. That single architectural distinction drives measurable differences in which sources get cited, how many brands get surfaced, and how often negative sentiment appears in the response.

Start With the Environment, Not Just the AI

Understanding the behavioral differences between these two platforms starts with understanding what each platform is embedded in.

Google AI Overviews appear within a search results page that already contains Shopping carousels, merchant product listings, local results, and organic links. A user who sees an AIO response has immediate access to purchase options below it. The AI can gesture toward a retailer, cite their domain, reference their pricing, surface their brand, and the SERP infrastructure does the rest.

ChatGPT has none of that. The response is the experience. If a user wants to act on a ChatGPT recommendation, the AI needs to provide enough evaluative context to justify the action. There's no carousel to fall back on. No organic listing to validate the pick. The AI is operating without a net, and the data shows it responds accordingly.

This isn't a flaw in either platform. It's by design. But it means brands need to understand not just whether they're being cited by AI, but where that citation is appearing and what the platform is being asked to do on its own.

AIO Cites Retailers Directly at Twice the Rate of ChatGPT

The most direct expression of this architectural difference: where citations actually go.

In Google AI Overviews, 30% of transactional citations reference a major retailer domain directly. In ChatGPT, that figure drops to 15%. Same purchase-intent query. Half the direct retailer presence.

The gap reflects the division of labor on each platform. AIO doesn't need to do as much evaluative work before pointing to a retailer because the SERP around it provides the commercial context. ChatGPT, operating without that context, routes citations differently before it arrives at a brand recommendation.

For retailers, this has a concrete implication. Being cited in AIO on transactional queries is a different kind of win than being cited in ChatGPT. AIO citation puts you on a page where the user is already in purchase mode. ChatGPT citation puts you in a response that still has more work to do before the user acts.

AIO Leans on Social Proof. ChatGPT Doesn't.

One of the more striking findings in the data is how differently each platform uses social and community content to anchor purchase recommendations.

YouTube and Facebook together account for nearly 13% of AIO's transactional citations. ChatGPT surfaces that same category at just 3%. A 4x gap. When Google's AI wants to validate a purchase recommendation, it reaches for peer content: video reviews, community discussions, social proof from real users. ChatGPT largely doesn't follow the same pattern.

This reflects a broader dynamic in how AIO handles the consideration layer. Where ChatGPT needs to do its own evaluative work in the text of the response, AIO can point users toward community content that carries that evaluation implicitly. A YouTube review, a community discussion, a video comparison — these are the validation signals AIO leans on. ChatGPT builds its own.

For brands, this matters beyond citation strategy. If your category's purchase journey is anchored in peer validation, and most retail categories are, your presence in social and video content isn't just a community play. It's an AIO citation surface.

ChatGPT Adds a Verification Layer Before It Recommends

Where AIO routes transactional citations toward retailers and social content, ChatGPT takes a different path. It goes to editorial and financial sources first.

Four of the top six most-cited domains in ChatGPT's transactional responses are editorial or financial sources: review outlets, deal-analysis sites, financial comparison platforms. In AIO, four of the top six are retailers. ChatGPT is adding a verification step that AIO largely doesn't need, because AIO's SERP already provides it through organic results and Shopping units.

The implication for brands is significant. A brand that isn't referenced by the editorial and financial sources ChatGPT trusts may be getting filtered out before the recommendation is made. The citation isn't just about whether your domain appears. It's about whether the sources ChatGPT relies on to validate purchases are already vouching for you.

ChatGPT Surfaces Wider Competitive Sets

ChatGPT surfaces an average of 7.5 brands per transactional response. AIO surfaces 6.1. That gap compounds across the buyer journey.

More brands per response means more options presented before a decision gets made. For any individual brand, it means the consideration set is wider and the path from AI response to purchase action is longer and more competitive on ChatGPT than on AIO.

This pattern is consistent with ChatGPT's role as a stand-alone evaluative layer. Without the SERP infrastructure to narrow the field, ChatGPT presents the user with more options and more context, letting the response do the comparison work that a SERP might distribute across multiple surfaces.

ChatGPT Is More Willing to Surface Negative Sentiment

Negative brand mentions in ChatGPT's transactional responses run at nearly double the rate of AIO's. 0.7% vs. 0.4%.

The absolute numbers are small. But the pattern matters. When ChatGPT is the only thing on the page, it bears full responsibility for a complete and balanced answer. That means it's more willing to surface reasons not to choose a brand, including compatibility issues, price concerns, and product limitations, as part of making its response useful. AIO, operating within a SERP that gives users more ways to evaluate on their own, applies a lighter editorial hand.

The practical implication: brands with product or experience weaknesses that are well-documented in editorial and review sources face more exposure in ChatGPT's transactional responses than in AIO's. Monitoring sentiment in AI responses isn't just a brand exercise. It's a transactional visibility issue.

What Marketers Need to Know

The behavior difference is architectural, not algorithmic. AIO and ChatGPT are both trying to answer the same question. The path they take depends on what's around them. Understanding that distinction is the starting point for any AI citation strategy in retail.

Social and video presence is a transactional citation surface on AIO. AIO's 4x higher rate of social and community citations on transactional queries means peer content, including YouTube reviews, community discussions, and video comparisons, is doing citation work in the purchase journey. Brands that don't show up in that content layer are absent at a critical moment.

ChatGPT's verification layer is the editorial web. If the review outlets, comparison sites, and financial sources ChatGPT trusts aren't vouching for your brand, you may be getting filtered before the recommendation is made. Visibility in those sources isn't just an SEO play. It's a ChatGPT transactional citation play.

The good news: the foundation is the same across both platforms. Authoritative content. Trusted source signals. Credibility at scale. The inputs that drive citation visibility on AIO are the same inputs that drive it on ChatGPT. They're just weighted and expressed differently depending on the environment. Brands that build that foundation don't need separate strategies for each platform. They need the visibility to see how each platform is interpreting what they've already built.

Technical Methodology

Key Takeaways

Google cites YouTube broadly. ChatGPT is selective. What that means for your video and AI search strategy.

You'd expect Google to favor YouTube — it's a Google property. But when we analyzed how each AI engine actually uses YouTube as a citation source, the story isn't just about volume. It's about editorial intent.

Google AI Overviews surfaces YouTube across an enormous range of queries — roughly 30x more than ChatGPT in absolute volume. But ChatGPT is far more deliberate about when and why it cites it. That selectivity reveals something important: these two engines have fundamentally different theories about what YouTube is for.

The implications for brands go beyond content creation. Before deciding whether to build a YouTube presence, the smarter move is to understand what AI is already citing for your category — and who owns it. A single video you don't control can shape what an AI engine says about your brand across thousands of queries.

We used BrightEdge AI Hypercube™ to analyze YouTube citation patterns across millions of prompts in Google AI Overviews and ChatGPT. Here's what we found.

Data Collected

Using BrightEdge AI Hypercube™, we analyzed:

Key Finding

Google AI Overviews and ChatGPT both cite YouTube — but for fundamentally different reasons, at different stages of the user journey, and for different types of content. Google uses YouTube broadly as a general authority source. ChatGPT uses YouTube selectively, concentrating its citations in two specific use cases: instructional how-to content and entertainment/streaming discovery.

The gap in absolute volume is striking — Google surfaces YouTube in roughly 30x more queries than ChatGPT. But the intent profile of ChatGPT's citations is sharper and more deliberate. Understanding that difference is the starting point for any YouTube strategy in the context of AI search.

The Scale Gap — and Why It Matters

The volume difference between the two engines is significant. Google AI Overviews operates across a far larger query surface and cites YouTube in millions of responses. ChatGPT's YouTube citations, by comparison, are concentrated and purposeful.

This distinction matters for strategy: if you're optimizing for Google AIO, you're trying to be relevant across a broad informational landscape. If you're optimizing for ChatGPT, you're competing for a smaller but more deliberate citation set — which means the bar for what gets cited is higher.

The brands that win on both engines are the ones with YouTube content that is both broad enough to surface across Google's wide citation surface and specific enough to clear ChatGPT's higher threshold.

ChatGPT Sees YouTube as a How-To Library

The most significant single finding in this analysis: 60% of ChatGPT's YouTube-cited queries are instructional — how-to content, step-by-step guides, skill-building queries. Google AI Overviews? Only 22%. ChatGPT is nearly 3x more likely to cite YouTube for instructional content.

For ChatGPT, YouTube isn't a general information source — it's specifically where it sends users to learn something. When someone asks ChatGPT how to build a fence, learn sign language, solve a Rubik's cube, or set up a Gmail account, it reaches for YouTube. That behavior is consistent and predictable across categories.

Google AIO distributes its YouTube citations much more broadly — across general informational queries, topic explainers, cultural content, and reference material that has nothing to do with step-by-step instruction. The how-to use case is important to AIO, but it's one of many.

What This Means

• For ChatGPT visibility, instructional video content is the primary entry point. If your category has significant how-to search volume, find out what videos ChatGPT is already citing before you decide whether to build or partner.
• For Google AIO, topic authority matters more than format. AIO will cite YouTube across a much wider range of content types — the question is whether your content, or content in your category, has the authority signals AIO looks for.
• A YouTube strategy built only around tutorials will perform well in ChatGPT but will capture only a fraction of the AIO opportunity.

ChatGPT Is Also an AI-Powered Streaming Guide

The second major use case where ChatGPT concentrates its YouTube citations: entertainment and streaming discovery. When users ask where to watch something — a show, a sporting event, a live broadcast — ChatGPT frequently surfaces YouTube as a destination alongside traditional streaming platforms.

The data shows this clearly: “where to watch” queries see ChatGPT citing YouTube nearly 7x more often than Google AI Overviews. Entertainment and media queries overall show ChatGPT at 2.5x higher citation frequency than AIO.

In this context, ChatGPT is functioning like a modern cable guide — positioning YouTube in a lineup alongside Netflix, Hulu, Apple TV+, and Amazon Prime Video. It treats YouTube TV as a legitimate streaming platform in its own right, with that co-citation appearing nearly 7x more often in ChatGPT than in Google AIO.

Google AIO largely doesn't play this role. When users ask AIO where to watch something, the response pattern is different — it tends to point to dedicated streaming platforms rather than positioning YouTube as a discovery destination.

What This Means

• For brands in entertainment, sports, live events, or any category with “where to watch” search volume: ChatGPT is the AI discovery layer you need to be present in.
• YouTube TV presence and YouTube channel visibility are directly relevant to how ChatGPT answers streaming and entertainment queries.
• If your content has any video distribution component, ChatGPT’s streaming-guide behavior makes YouTube citation a reachable goal — provided the right content exists to be cited.

Google AIO Owns the Purchase Journey

Where ChatGPT pulls back from YouTube, Google AI Overviews leans in: the research and consideration phase of the buying journey.

Review and comparison queries — “best,” “vs,” “top,” “compare” — see Google AIO citing YouTube 2.5x more than ChatGPT. Consideration-intent queries broadly run 2x higher in AIO. Post-purchase intent queries also skew toward AIO.

When someone is actively evaluating a product, comparing options, or deciding what to buy, Google pulls YouTube into the answer. A product review video, a side-by-side comparison, an “is it worth it” breakdown — these are the formats AIO reaches for at the consideration stage. ChatGPT, for those same types of queries, mostly doesn’t.

This is a meaningful distinction for brand strategy: YouTube content that performs well in the purchase journey — review-style, comparison-style, evaluative — has a clearer path to AIO citations than to ChatGPT. And given AIO’s position at a high-volume point in the consumer research process, that’s a high-value citation surface.

What This Means

• Product review content, unboxing videos, “is it worth it” formats, and comparison-style videos are the highest-leverage YouTube content types for Google AIO visibility.
• Brands that don’t own YouTube presence in their category’s consideration-stage queries may be ceding that AIO citation surface to independent reviewers, competitors, or creators.
• ChatGPT is largely not the channel for purchase-journey YouTube citations — AIO is where that battle is fought.

The Strategic Framework: Audit Before You Build

The instinct when seeing this data is to say “we need more YouTube content.” That may be right. But the more important first step is understanding what YouTube content AI is already citing for your category — and who owns it.

We’ve seen cases where a single YouTube video, not owned by the brand, was controlling what an AI engine said about that brand across thousands of queries. That’s a risk if the framing isn’t favorable. It’s also an opportunity — if you know it’s happening and can act on it.

The strategic question isn’t “should we make more YouTube content?” It’s: which videos is AI already pulling for my category’s key queries, who owns them, and is there a faster path to AI citation through partnership than through production?

The Audit-First Approach

• Identify which YouTube videos are being cited by each AI engine for your category’s highest-value queries
• Determine whether those citations are from owned content, competitor content, or independent creators
• Assess whether influential creators in your category already have AI’s trust on topics you need to own
• Map the gap: is this a content creation problem or a content partnership problem?

What Marketers Need to Know

1. The real strategic question isn’t “should we make more YouTube content?”

It’s: what is AI already citing for your category, and who owns it? A single video you don’t control can shape what AI says about your brand at scale. That’s both a risk and an opportunity — but only if you know it’s happening.

2. Google and ChatGPT use YouTube for completely different jobs.

Google cites YouTube broadly across millions of queries as a general authority signal. ChatGPT is selective — concentrating citations in instructional content and entertainment discovery. A YouTube strategy that serves one engine may be largely invisible to the other.

3. For ChatGPT, instructional content is the entry point.

60% of ChatGPT’s YouTube citations come from how-to queries. If your category has instructional search volume, find out what videos ChatGPT is currently pulling before you decide whether to build or partner with a creator who already has that authority.

4. For Google AIO, YouTube citations run deepest in the purchase journey.

Review, comparison, and consideration-intent queries are where AIO leans on YouTube most. That’s where owned or partnered video content carries the highest strategic value — and where ceding that ground to independent reviewers creates the most risk.

5. Partnership is often the faster path.

Creators who already have AI’s trust in a category represent an alternative to building from scratch. Getting your brand into the conversation through an established channel may generate AI citations faster than building a new one — and is particularly relevant in categories where independent creators dominate the current citation landscape.

Technical Methodology

Key Takeaways