The conventional wisdom said AI search would compress the funnel and shift citations toward third-party reviewers and aggregators in the consideration phase. The data tells a different story. Across eight industries on Google AI Overviews and ChatGPT, the middle of the funnel still represents meaningful volume, and brand-owned content takes the dominant share of citations there.

BrightEdge AI Hyper Cube analysis across Google AI Overviews and ChatGPT shows that the consideration stage of the funnel is alive, measurable, and varies dramatically by industry. The size of the middle ranges from 4% to 26% of AI search demand depending on the category and the engine. Inside that middle, brand-owned pages account for 42% to 79% of citations across every industry studied, while review and comparison aggregators account for 1% to 7%. The opportunity for brands is on their own domains.

This is the second installment in our funnel-shape research series. In the prior installment we examined how the four classical query intent categories survived into AI search, each reshaped to fit the medium. That analysis mapped consideration-stage queries to the classical commercial intent bucket. This installment looks at the same data through the lens of the consumer journey: top, middle, and bottom of the funnel. The terminology shifts from "commercial intent" to "consideration stage" because the questions we are answering this time are about consumer-journey shape, not classical SEO intent taxonomy. The underlying data is consistent with the prior piece.

We analyzed the full prompt universe across eight industries on both engines: B2B, Ecommerce, Education, Entertainment, Finance, Healthcare, Insurance, and Travel. Each prompt was classified by the BrightEdge Generative Parser into a funnel stage. Citations from consideration-stage prompts were categorized by source type. The findings are directly relevant to any brand planning content strategy for AI search across multiple industries.

Data Collected

Key Finding

The middle of the AI search funnel is real volume in every industry, and brand-owned content owns it. Volume-weighted, the consideration stage represents between 4% and 26% of AI search demand across the eight industries studied. Travel, Ecommerce, B2B, and Finance show the largest middles. Healthcare and Entertainment show the smallest. Inside that consideration stage, brand-owned commercial pages take between 42% and 79% of all citations across every industry on both engines. Review and comparison aggregators, the source type marketers most often assume dominates the research phase, take between 1% and 7% of citations in most categories. The implication for marketers is that the buying-guide content, category explainers, and comparison pages on a brand's own domain are the highest-leverage AI search assets in the middle of the funnel. Outsourcing the consideration phase to third-party reviewers leaves the dominant citation channel uncovered.

The Size of the Middle, by Industry

Volume-weighted share of AI search demand classified as consideration stage:

Two patterns stand out. First, the size of the middle of the funnel varies by a factor of six between the largest category and the smallest. Marketers who assume the consumer journey looks the same across industries are missing where the opportunity actually concentrates. Second, AI Overviews consistently shows a larger consideration stage than ChatGPT in commercial categories. In Travel, Ecommerce, B2B, Finance, and Insurance, AIO's middle is meaningfully bigger than ChatGPT's. In Education and Entertainment, the two engines look roughly the same. The pattern suggests AIO is woven into the buying journey in commercial categories in a way ChatGPT is not, despite the popular narrative that consumers have moved their research behavior to conversational AI.

What a Consideration Prompt Looks Like

Consideration prompts capture users who are comparing options or evaluating a category without committing to a specific brand or product. Examples drawn from the data, generalized for clarity:

In Travel: "best beach vacations for families," "top all-inclusive resorts in Mexico," "cheapest time to fly to Europe."

In Ecommerce: "best treadmill for home use," "treadmill vs elliptical for cardio," "best mattress for back pain."

In B2B: "small business CRM software," "best project management tools for remote teams," "top cloud storage providers for enterprise."

In Finance: "best high yield savings accounts," "Roth IRA vs traditional IRA," "top robo-advisors."

In Insurance: "term vs whole life insurance," "best homeowners insurance companies," "cheapest car insurance for new drivers."

These queries differ from branded queries (which name a specific product or company) and from transactional queries (which signal readiness to act). The defining characteristic is comparison and evaluation. The user is figuring out what to want, not which one to click.

Brand-Owned Content Dominates Consideration-Stage Citations

Across every industry studied and both engines, brand-owned commercial pages take the largest share of citations in the consideration stage. The range is 42% on the low end (Healthcare, where authority sites take a meaningful slice) to 79% on the high end (Travel ChatGPT, where the engine routes consideration queries heavily to brand domains and bypasses online travel agencies). Most industries land between 50% and 70%.

This finding pushes back on a widely held assumption in the AEO and GEO community. The assumption was that AI engines, when faced with a comparison query, would lean on third-party reviewers and aggregators to make the recommendation. The data shows the opposite. The brand's own buying guide, category explainer, or comparison page is more often the cited source than a review aggregator.

The pattern is consistent across engines, with one nuance. AIO concentrates citations across a small number of brand-owned domains. ChatGPT distributes citations across a wider set of brand-owned domains. The dominance of brand-owned content holds in both cases, but the competitive dynamics are different. On AIO, winning a consideration-stage citation in a given category means displacing a small number of established players. On ChatGPT, winning a citation is more accessible, but the citation share per win is smaller.

Review and Comparison Aggregators Are Not the Dominant Source

The source type that conventional AEO wisdom positioned as the natural winner of the consideration stage, review and comparison aggregators, accounts for between 1% and 7% of citations in most industries on both engines. The two exceptions are B2B on ChatGPT, where software review and comparison sites take a slightly larger share, and Finance on AIO, where financial product comparison sites cluster around the high end of the range. Even in those exceptions, brand-owned pages still take more citations than aggregators.

This does not mean third-party reviews are irrelevant. They influence the brand recommendations AI engines surface and they remain important for trust signals. But the citation slot, the actual source AI engines link to in the consideration stage, more often belongs to a brand's own domain. Marketers who have built their AI search strategy primarily around earning third-party reviewer mentions are competing for a small share of the citation channel.

Citation Concentration: AIO Concentrates, ChatGPT Distributes

Citation concentration in the consideration stage differs substantially between engines. On Google AI Overviews, a small number of domains accounts for the majority of citations in any given industry. On ChatGPT, citations spread across a much larger set of domains for the same industries.

The pattern means winning consideration-stage citations on AIO requires going deeper on a smaller number of pages within a category. The competitive set is narrow. Once a brand earns a citation slot, it tends to hold it across many related prompts. ChatGPT is the opposite. The citation pool is more democratic. Breadth of topical coverage, distinctive perspectives on a category, and content depth across many comparison angles matter more than dominance on a single page.

For content strategy, this means the optimization approach differs by engine. On AIO, the priority is identifying the small number of pages that win the highest-volume consideration queries in a category and concentrating optimization investment there. On ChatGPT, the priority is breadth, coverage across the full comparison landscape, and content depth that signals authority across many sub-topics within a category.

A Note on Google AI Overviews and When They Trigger

Google AI Overviews do not appear on every search. AIO is triggered only when Google decides an AI Overview is the right response format for a given query. Many consideration-stage searches return a traditional organic results page with no AIO at all. The analysis in this study measures the subset of consideration queries where Google has chosen to deploy an AIO.

This caveat actually strengthens the finding rather than weakening it. Even on the consideration queries Google has decided merit an AI Overview, the citation slots are not spreading across third-party reviewers and aggregators. They are concentrating on brand-owned pages. Whatever combination of signals Google uses to decide when to deploy AIO and what to cite inside it, the result is that brand-owned content is the dominant beneficiary in the consideration stage.

ChatGPT does not have an equivalent trigger condition. Every prompt receives a response. The full consideration-stage volume on ChatGPT is measured directly. The fact that brand-owned content dominates on both surfaces, despite the very different mechanics of how AIO and ChatGPT decide what to cite, reinforces the strength of the underlying pattern.

Industry-Specific Patterns

Healthcare. Healthcare authority sites (major medical centers, government health agencies, established medical reference sites) take a larger share of citations than in any other industry, between 26% on AIO and 36% on ChatGPT. Even so, brand-owned commercial pages still take the largest single share. The takeaway for healthcare marketers is that competing for citation slots means competing against highly credentialed authority sources, which makes E-E-A-T signals (expertise, experience, authoritativeness, trust) even more important in this category than elsewhere.

Travel. Travel shows the most divergent engine behavior. On AIO, online travel agencies and booking aggregators take a meaningful slice of consideration citations (around 24%). On ChatGPT, the engine bypasses OTAs and routes consideration citations directly to brand-owned destinations (79% brand-owned). For travel brands, this means a ChatGPT optimization strategy that targets brand-owned travel content can win significant citation share, while an AIO strategy needs to plan for OTA competition in the citation slot.

B2B. B2B shows the cleanest gap between AIO's larger middle of the funnel (22%) and ChatGPT's smaller middle (9%). The implication is that B2B buyers are using AIO for category exploration more than they are using ChatGPT for the same purpose, at least in the consideration stage. Software review aggregators have a slightly more prominent role here than in other categories, but brand-owned product pages and buyer's guides still take the largest share.

Education and Entertainment. These are the only two industries where ChatGPT's middle of the funnel is larger than or equal to AIO's. Both categories also show meaningful citation share for video platforms (10% to 17% on AIO) and UGC sources (10% to 13% on both engines). The pattern suggests that for educational and entertainment decisions, users are pulling in more diverse source types than in commercial categories.

What Marketers Need to Know

The middle of the funnel is real volume in AI search. The size varies by industry and by engine, ranging from 4% to 26% of total demand. In Travel, Ecommerce, B2B, and Finance, the consideration stage represents a meaningful share of AI search demand on both engines and should be a primary focus for content strategy.

Your own content is the opportunity. Brand-owned commercial pages take 42% to 79% of consideration-stage citations across every industry studied. The category guides, comparison pages, and buying guides on your own domain are doing the work. Investing in this content is more leveraged than chasing third-party reviewer placements.

Do not outsource the middle to third parties. Review and comparison aggregators take 1% to 7% of consideration citations in most categories. They remain important for trust signals and indirect influence on what AI engines recommend, but the citation channel itself belongs to brand-owned content.

Optimize for both engines differently. AIO concentrates citations across a small number of pages. ChatGPT distributes citations across a wider set. The same brand-owned content strategy serves both engines, but the tactical priorities differ. On AIO, win the small number of pages that own the highest-volume consideration queries. On ChatGPT, build breadth and topical depth across the full comparison landscape.

Audit your consideration coverage by industry. Some industries have much larger middles than others. If you compete in Travel, Ecommerce, B2B, or Finance, the consideration stage deserves significant share of your AI search investment. If you compete in Healthcare, Education, or Entertainment, the middle is smaller, but the source-type dynamics in those categories require category-specific strategy (authority signals in Healthcare, video and UGC presence in Education and Entertainment).

Expect engine-specific behavior, not engine-specific intent. The underlying user behavior in the consideration stage is the same across engines. The way each engine surfaces and cites sources for that behavior differs. A unified content strategy organized around the consumer journey, with execution tuned for each engine's citation dynamics, is more durable than separate engine-specific playbooks.

Technical Methodology

Key Takeaways