How Agent Edge for CloudFront wins your brand visibility in AI search, without slowing your site or risking the SEO you have already earned.
Originally presented on September 16, 2026, this on-demand session explores how AI-agent traffic is changing the demands on enterprise websites and what teams can do to prepare at the infrastructure layer.
AI agents are becoming a new source of traffic, discovery, and brand visibility, creating new challenges for enterprise websites at the edge. In fact, BrightEdge research finds that 97% of the web's top 20,000 sites aren't yet set up to serve AI agents optimized content.
Join AWS and BrightEdge for a 30-minute look at Agent Edge for CloudFront, a new solution designed to help enterprises detect AI-agent traffic at the edge while improving performance and making content easier for AI agents to access and understand without changing the experience for humans or traditional search engines.
Key Takeaways
.png)
.jpg)
BrightEdge Google Search Console Reporting
Google Search Console (GSC) is the most precise research database for Search, showing you the exact words people type into Google to find you. That data comes straight from the search engine that handles close to 90 percent of the world's search queries and every impression and click is triggered by a live user, not based on a synthetic model or panel data.
BrightEdge GSC Reporting takes that same data, makes it much easier to explore than with Google’s native UI, and adds several insights to make that data actionable. Specifically, it removes the limit Google's interface puts on how many keywords and pages you can see. It keeps a longer look back than Google's own dashboard allows and saves your filters, so you are not rebuilding them every time you open a report. BrightEdge GSC Reporting also connects the dots for you: why clicks dropped even as visibility climbed, which specific keywords and pages are behind any swing and where an AI Overview is quietly taking the click before someone reaches your page. And the best part is that all you need to do is connect Google GSC to your BrightEdge account which takes less than 1 minute.
GSC 1.png)
Complete Search Coverage
Native Google Search Console limits the keywords and pages it shows in its interface to 1000, so medium and large sites hit that ceiling immediately. BrightEdge GSC Reporting removes that limit and shows every keyword and page tied to your organic performance. You see the long tail of demand Google's own interface hides, and you catch swing in pages ranked below the cutoff before a drop in traffic looks like a mystery.
AI Impact Diagnosis
A common finding is that impressions climb while clicks fall. It looks like Native Google Search Console is telling two different stories at once. BrightEdge GSC Reporting filters your keyword and page performance by SERP feature, including AI Overview presence. This isolates the exact queries affected by AI. Each keyword also carries its exact (not broad) search volume next to the click and impression numbers, giving every drop real demand context. Now you can tell a real ranking problem apart from an AI Overview taking the click. You know whether the fix is better content and ranking or whether no ranking change would bring the click back.
Performance Breakdown
BrightEdge GSC Reporting makes it easy to compare any two periods. That could be the last week against the week before, or the last quarter against the quarter before. It surfaces the keywords and pages behind the swing. You can split the view by branded and non-branded terms too, without rebuilding your filters every time you open a report. Every page links straight through to the keywords driving its clicks and impressions. That allows you to easily navigate from a high-level change to the specific queries causing it.
Custom Reporting
Native Google Search Console forgets every filter the moment you close the report. Keyword and URL filtering starts over each time. BrightEdge GSC Reporting remembers your branded terms, keyword groups and page groups. It applies them automatically to every report after that. You can group keywords by category. You can also build page groups from a bulk URL upload or an automatic page rule, such as everything under a specific subfolder. Once those groups exist, every future report opens with them already applied.
Ready to know why your traffic moved?
Native Google Search Console limits the keywords and pages it shows in its interface. Larger sites with more organic traffic hit that ceiling the fastest. GSC Reporting removes that limit and shows every keyword and page tied to organic performance, with the same filtering and comparison tools available across the full data set.
A drop in clicks alongside rising impressions is a common pattern. It often means an AI Overview or another SERP feature is now appearing on those queries and "stealing” the click. BrightEdge GSC Reporting isolates keywords by SERP feature, including AI Overview presence, so you can confirm the cause before assuming a ranking loss.
Native Google Search Console requires a manual filter to be rebuilt for branded and non-branded terms every time a report runs. BrightEdge GSC Reporting classifies brand terms once and applies that split automatically across every report, chart and comparison going forward.
Native Google Search Console requires manually selecting two custom date ranges to compare. It has no saved year-over-year preset. GSC Reporting lets you compare any week, month or quarter against the same period a year earlier, in a couple of clicks. That helps you spot seasonal patterns a shorter view would miss.
BrightEdge GSC Reporting supports page groups built from a bulk URL upload or an automatic page rule, so you can view Search Console performance for a subfolder, template or content type without rebuilding a filter each time.
More than 700 BrightEdge customers already use GSC Reporting. It helps them catch the click loss AI Overviews are causing, an insight Google's own tool will not surface. It runs inside the BrightEdge platform your team already uses for keyword, content and AI search reporting. No separate login or export is required.
Last week's edition traced a single publisher's rise in Google AI Overviews back to a shift in what people were asking. This one goes a level deeper, into who each engine treats as an authority once the question is asked. And into how much room that leaves anyone else.
We pulled the cited-domain lists for healthcare and finance, the two categories Google's quality guidance holds to the highest standard. Both engines, in AI Hyper Cube. They produce different lists in the same category, in the same week.
YMYL stands for Your Money or Your Life. It is Google's own term, set out in its Search Quality Rater Guidelines. It covers topics where low-quality content could plausibly harm a person's health, financial stability, safety or wellbeing. Pages in those categories are held to a higher standard than the rest of the web.
| YMYL category | What it covers | Measured in this edition |
| Health and medical | Diseases, treatments, symptoms, mental health | Yes. Both engines |
| Financial advice | Investments, taxes, loans, banking | Yes. Both engines |
| Legal guidance | Divorce, custody, rights, obligations | Not yet |
| News and civics | Current events, government policy, voting information | Not yet |
Source: Category definitions from Google's Search Quality Rater Guidelines. Coverage from BrightEdge AI Hyper Cube, week of 23 August 2026, United States.
We measured the first two. Legal guidance and news and civics are not in this dataset, so nothing here should be read as a claim about them. Both are on the list to run next.
The reason the distinction matters is that the guidance describes a standard rather than a source list. It does not name which domains an engine should cite in a sensitive category, and the two engines we measured have reached different answers. AI Hyper Cube reports the cited-domain list per engine, so you can check your own category rather than infer it from the guidance.
In tracked healthcare prompts, every one of ChatGPT's five most-cited domains is a government agency or a non-profit hospital system. Across the top ten, that holds for nine. YouTube does not appear at all.
| Domain | Citation presence, share of tracked healthcare prompts |
| nih.gov | 43% |
| medlineplus.gov | 33% |
| mayoclinic.org | 27% |
| cdc.gov | 18% |
| clevelandclinic.org | 17% |
Source: BrightEdge Insights from AI Hyper Cube. Citation presence, week of 23 August 2026, United States.
If your category is safety-adjacent, that is the citation diet you compete against on this engine.
Two categories, two engines. For healthcare we took the five most-cited domains on each engine and their citation presence. Citation presence is the share of tracked prompts in a category where a domain appeared as a cited source. We also took the weekly healthcare series on Google, indexed so every source starts at the same point. For finance we took citation presence at each rank position within each engine's top five.
Each category is a separate prompt set. We compare figures within a category and engine, never across them. The prompt sets are different sizes, so comparing across them would compare the sets instead of the engines.
| Dimension | Detail |
| Capability | BrightEdge AI Hyper Cube, cited-domain and citations-trend views. AI Catalyst for the measurement guidance |
| Engines | Google AI Overviews, ChatGPT |
| Categories | Healthcare, finance. B2B technology as a non-sensitive control |
| Period | Week of 23 August 2026. Weekly healthcare series from 7 June |
| Metric | Citation presence, the share of tracked prompts in a category citing a given domain |
| Comparison basis | Within category and engine. Weekly series indexed to a common baseline |
| Geography | United States |
| Anonymization | Reported as shares and ratios only. No prompt or keyword counts |
What the data shows. Measured: ChatGPT's healthcare top five is entirely government and hospital domains, with YouTube absent. In the same category on Google AI Overviews, YouTube stepped to roughly 1.8 times its June level in the week of 19 July, and held. Clinical partners stayed near baseline.
What we take from it. Inferred: the two engines are building different definitions of what counts as an authoritative source. The difference is large enough to change what a marketer should build. The data measures the lists. The read on what to build from them is ours.
The same category on Google looks nothing like ChatGPT's. In the week of 19 July, YouTube's citation presence in healthcare AI Overviews stepped up sharply and held there. NIH drifted below its June level.
| Source | Presence, week of 7 June | Presence, week of 23 August |
| youtube.com | 100 | 158 |
| clevelandclinic.org | 100 | 108 |
| nih.gov | 100 | 83 |
Source: BrightEdge Insights from AI Hyper Cube. Google AI Overview citations, indexed so each series starts at 100 in the week of 7 June 2026, United States.
Brands with authoritative content still place. They now place alongside individual clinicians, museums and creators. One video surfacing in this set is a pathology museum explaining a condition with a specimen in hand. The asset was distinctive rather than institutional.
Finance shows the same disagreement in a different shape. On Google the leading source is cited in roughly half of all tracked prompts. The four below it are the largest sites in the category. On ChatGPT the same list is close to level.
| Rank position | Google AI Overviews | ChatGPT |
| 1 | 51% | 19% |
| 2 | 29% | 15% |
| 3 | 19% | 14% |
| 4 | 19% | 14% |
| 5 | 18% | 11% |
Source: BrightEdge Insights from AI Hyper Cube. Citation presence as a share of tracked finance prompts, week of 23 August 2026, United States.
simplywall.st holds the fifth slot on ChatGPT. It is not a large publisher. It publishes community valuation narratives, where named contributors argue a fair value for a stock against analyst positions. No larger competitor in that list holds that view of a ticker.
Divide the leading source's presence by the fifth and you get a number worth running on your own category. Finance is 2.8x on Google and 1.7x on ChatGPT. A steep curve means an entrenched incumbent. A flat one means the category is shared.
We added B2B technology as a control, because Google's strictest guidance does not cover it. The engine split appears there too, which suggests it tracks the engine rather than the sensitivity of the subject.
| Category | ChatGPT top five | Google AI Overviews top five |
| Healthcare | All government or hospital. No YouTube | YouTube first, above every clinical source |
| Finance | Spread across five sources, one specialist | One source at roughly half of prompts |
| B2B technology | nist.gov first, cisa.gov fourth | YouTube first, at two thirds of prompts |
Source: BrightEdge Insights from AI Hyper Cube. Citation presence, week of 23 August 2026, United States.
Three categories is a pattern. A rule would need more. Run the same cut on your own category before you assume it behaves like any of these. Then compare it against your organic footprint in Data Cube X, to see whether the pages an engine cites are the ones already ranking.
1. Find the thing only you can say. Without a distinct point of view or data nobody else holds, selection is luck of the draw. simplywall.st takes a finance top-five slot on ChatGPT on community valuation narratives no larger competitor publishes. Establish what you or your audience know that the category cannot easily aggregate.
2. Read the citation list for your category, per engine. The winning source type is not a constant. Government and hospital domains take all five of ChatGPT's healthcare slots. Google AI Overviews puts YouTube above every clinical source in the same category. Pull the list before you decide what to build.
3. Check how much room the list leaves you. Divide the leading source's citation presence by the fifth. Finance runs 2.8x on Google and 1.7x on ChatGPT. Above three you are displacing an incumbent. Under two you are joining a rotation, which is a far cheaper fight.
4. Start with Google, and do not read format as a shortcut. Google AI Overviews carries the volume, so plan there first. Video and community formats reach places a long-form article does not, and that is a distribution decision. Foundational SEO remains the cost of entry to both surfaces.
What is YMYL content?
YMYL stands for Your Money or Your Life. It is Google's term for topics where low-quality content could harm a person's health, financial stability, safety or wellbeing. Google's Search Quality Rater Guidelines group them into health and medical, financial advice, legal guidance, and news and civics.
What are the YMYL categories?
Four. Health and medical, covering diseases, treatments and mental health. Financial advice, covering investments, taxes, loans and banking. Legal guidance, covering rights and obligations. News and civics, covering current events, government policy and voting information.
Do Google AI Overviews and ChatGPT cite the same sources in YMYL categories?
Not in this dataset. In healthcare, ChatGPT's five most-cited domains are government agencies and hospital systems, and YouTube does not appear. Google AI Overviews places YouTube above every clinical source in the same category and the same week.
Can a smaller site get cited in a YMYL category?
In finance, yes. Citation presence on ChatGPT is spread almost evenly across the top five, and simplywall.st holds a position on community valuation narratives no larger competitor publishes. Concentration is far higher on Google AI Overviews, where the leading source appears in roughly half of tracked prompts.
How do you find out which sources get cited in your category?
Pull the cited-domain list for your category on each engine and compare them. BrightEdge AI Hyper Cube reports citation presence per engine. AI Catalyst tracks how your own share moves week over week. Data Cube X shows whether the cited pages are the ones already ranking in organic search.
| Item | Detail |
| Data source | BrightEdge AI Hyper Cube, cited-domain and citations-trend views |
| Engines | Google AI Overviews, ChatGPT |
| Categories | Healthcare, finance, and B2B technology as a non-sensitive control |
| Period | Week of 23 August 2026. Weekly healthcare series from 7 June 2026 |
| Measurement | Citation presence, the share of tracked prompts in a category citing a given domain |
| Indexing | The weekly series is indexed so every source starts at 100 in the week of 7 June, showing movement against a common baseline rather than absolute levels |
| Comparison basis | Within category and engine only. Never across categories |
| Geography | United States |
| Anonymization | Shares and ratios only. No prompt counts, keyword counts, or figures that back-calculate to either |
| Limits | Three categories and one week, plus one weekly series. Directional, and not a claim about every vertical or every period |
| Finding | The number that proves it |
| ChatGPT treats institutions as the authority in health | All five of the top five are government or hospital. Nine of the top ten |
| Google AI Overviews treats the platform as the authority in the same category | YouTube stepped to 1.8x its June level in the week of 19 July and held |
| Google concentrates finance citations on one source | 51% presence at rank one, against 18% at rank five |
| ChatGPT spreads them | 19, 15, 14, 14 and 11 percent across the top five |
| A flat list is where a specialist site gets in | simplywall.st at rank five on ChatGPT in finance |
The downloadable edition carries all three charts, the full cited-domain lists, the B2B technology control, and the methodology. It is the version to send to your search and content leads.
Find out what counts as authority in your category. Request a demo.
Explore more in our Weekly AI Search Insights series.
Pull your category's cited-domain list in AI Hyper Cube. Track your own share week over week in AI Catalyst. Compare it against your organic footprint in Data Cube X.
Featured in the Forbes CIO newsletter, BrightEdge’s latest research reveals that 97% of websites remain unoptimized for AI agent discovery, with only 3% utilizing agent-friendly formats such as Markdown. BrightEdge founder and CEO Jim Yu described this paradigm shift as “the rise of almost like a second web,” where autonomous AI agents act as the primary audience navigating services and recommending brands to consumers. He emphasized that enterprises do not need to overhaul existing sites built for human search, but should instead deploy synchronized, AI-tailored experiences in parallel. In BrightEdge pilot implementations, serving optimized content to AI agents delivered a decisive competitive edge alongside a 90% reduction in site payload
Two weeks ago this series measured Reddit's expansion in search visibility and found a tracked footprint roughly 33 times larger than a year earlier, with Google AI Overviews carrying about 88% of Reddit's AI citation volume. That study established scale. This one asks a different question. When an engine cites Reddit, what did the person already know before they asked?
We used BrightEdge AI Hyper Cube to analyze the prompts behind Reddit citations across nine industries on Google AI Overviews and ChatGPT. In categories where people shop by name, a large share of those prompts already contain the brand that the answer goes on to return. Reddit is not being consulted to assemble a shortlist. It is being consulted about a name the shopper already has.
For each Reddit-cited prompt we checked whether the prompt text itself names a brand that appears in the answer. The pattern separates cleanly between categories where purchase decisions are brand-led and categories where they are not.
| Industry | Google AI Overviews | ChatGPT |
| Restaurants | 43.5% | 52.0% |
| Ecommerce | 36.4% | 49.6% |
| B2B | 36.1% | 51.6% |
| Insurance | 27.8% | 33.8% |
| Finance | 27.3% | 35.7% |
| Travel | 22.8% | 23.2% |
| Entertainment | 21.5% | 42.6% |
| Education | 9.0% | 10.5% |
| Healthcare | 7.1% | 4.7% |
These figures use exact brand matching against the prompt text and represent a floor rather than a midpoint. Shoppers frequently use short forms of a brand name that a strict match discards.
We measured, for each industry and each engine, the share of Reddit-cited prompts in which the prompt names a brand returned by the answer, the number of distinct brands returned per Reddit-cited answer, and the distribution of cited prompts by intent stage. Brand presence was classified by exact word-bounded matching between prompt text and the brands recorded in the answer. Intent stage was taken from the classification attached to each prompt. All findings are reported as proportions within an industry and engine, so that differences between engines remain the unit of analysis rather than differences in the size of either prompt set.
| Data Point | Description |
| Engines analyzed | Google AI Overviews, ChatGPT |
| Platform analyzed | |
| Industries | B2B, ecommerce, education, entertainment, finance, healthcare, insurance, restaurants, travel |
| Period | 17 May 2026 through 2 August 2026 |
| Metrics | Share of cited prompts naming a returned brand; distinct brands returned per answer; distribution of cited prompts by intent stage |
| Comparison basis | Proportions within industry and engine |
| Anonymization | Findings reported as shares, rates and proportions, not raw prompt or citation counts |
Key Finding: This is not a Google phenomenon. ChatGPT cites Reddit far less often in absolute terms, but its Reddit citations are more concentrated on brand-aware prompts, in every industry analyzed except healthcare. Google carries the volume. ChatGPT carries the intent. Both point at the same stage of the buying process.
The intent classification supports the same conclusion from an independent direction. In the categories where brand-aware prompts are most common, consideration-stage prompts account for the largest share of Reddit's cited prompts.
| Industry | Share of Reddit-cited prompts at consideration stage, Google AI Overviews |
| Insurance | 41.8% |
| Ecommerce | 36.1% |
| Finance | 35.6% |
| Restaurants | 33.9% |
| Education | 29.9% |
| B2B | 28.7% |
| Travel | 20.4% |
| Entertainment | 9.9% |
| Healthcare | 4.7% |
Two independently classified measures agree. Prompts show shoppers arriving with a brand, and intent classification shows them arriving at the point of decision.
Healthcare is the lowest industry on both measures. Only 7.1% of its Reddit-cited prompts on Google name a brand, and only 4.7% sit at the consideration stage. People arrive at a healthcare question with a symptom rather than a company, and Reddit is retrieved for something other than vetting.
The vetting behavior appears precisely where purchase decisions are brand-led. It is a property of how people shop in a category, not a property of Reddit.
Reddit-cited answers return an average of 4.32 brands on Google AI Overviews and 4.56 on ChatGPT, with a median of four on both. Answers returning five or more brands account for 37.7% on Google and 41.8% on ChatGPT.
The influence is equivalent. A thread that no one on a marketing team has read is contributing to the set of names a buyer sees, and it requires no audience to do so. It only requires citation.
Among Google's Reddit-cited prompts in restaurants, 35.1% contain the word reviews. No other industry exceeds 3%. On ChatGPT the same figure for restaurants is 1.8%.
For operators with seasonal locations, holiday dining programs or catering services, this is a fourth-quarter exposure. The review query is being answered, and Reddit threads are answering it.
Audit what AI already cites for your brand terms before deciding whether to participate. The threads currently cited are the verdict your buyer receives. Where a brand has a genuine community presence, participation is natural. Where it does not, forced participation reads as marketing.
Do not budget Reddit as reach. Reddit visibility is not upper-funnel awareness. It operates between a shopper knowing a brand and a shopper choosing it, which makes it a demand-capture concern rather than a demand-generation one.
Do not treat this as a Google-only problem. Google cites Reddit more often, but ChatGPT's Reddit citations skew harder toward brand-aware shoppers in eight of nine industries.
Weight the effort to your category. Brand-aware prompts run above 36% in restaurants, ecommerce and B2B on Google and above 49% on ChatGPT. In healthcare and education they run under 11% on both.
Publishing more content does not address this. The cited asset is a third-party thread that no brand controls editorially. The corrective action is monitoring and response, not production.
| Parameter | Detail |
| Data Source | BrightEdge AI Hyper Cube |
| Engines Analyzed | Google AI Overviews, ChatGPT |
| Platform Analyzed | |
| Industries | Nine, as listed above |
| Period | 17 May 2026 through 2 August 2026 |
| Measurement | Cited prompt classification by brand presence, brand count returned, and intent stage |
| Brand matching | Exact word-bounded match between prompt text and brands recorded in the answer. Reported figures represent a lower bound |
| Comparison Basis | Proportions within industry and engine |
| Anonymization | Findings reported as shares, rates and proportions, not raw prompt or citation counts |
| Finding | Detail |
| Shoppers arrive with the brand already in hand | At least 36% of Reddit-cited prompts in restaurants, ecommerce and B2B on Google already name a brand the answer returns. On ChatGPT the same industries run above 49% |
| Reddit sits at the vetting stage | Consideration-stage prompts account for 41.8% of Reddit citations in insurance and 36.1% in ecommerce on Google AI Overviews |
| Both engines, not just Google | ChatGPT's Reddit citations skew more brand-aware than Google's in eight of nine industries, despite far lower citation volume |
| The shortlist is the same size either way | Reddit-cited answers return a median of four brands on both engines |
| Healthcare is the exception that proves the rule | Healthcare is lowest on both measures, at 7.1% brand-aware and 4.7% consideration stage, because people arrive with a symptom rather than a company |
| Restaurants face a fourth-quarter exposure | 35.1% of Google's Reddit-cited prompts in restaurants contain the word reviews, against under 3% in every other industry |
BrightEdge AI Agent Edge
AI agents such as ChatGPT, Perplexity and Claude now research and act on behalf of your customers, often before those customers ever visit your website. This issue is that AI agents have a hard time understanding and navigating through websites that were built with humans only in mind. BrightEdge Agent Edge is built to solve this: to make it easy for AI agents to understand your content so that when they recommend a brand, they recommend yours. And the best part is that it can do this in minutes, with minimal changes to your website.
"BrightEdge made it easy for us to measure our agent traffic and to establish a baseline. Together, we quickly implemented a solution to direct agents to AI-optimized markdown. Immediately, we saw increased traffic from high-value agents. Best of all, we did this with no change to our content creation and publishing workflows."
Measurable Impact
Sites using Agent Edge have seen agent traffic double or triple within weeks, with lifts in AI referral traffic as high as 30 percent. Agent Edge’s dashboard tracks the impact on your own pages, showing agent visit frequency before and after the change, along with which AI agents are visiting and which of your pages they favor. That traffic connects to your AI visibility, leads and revenue, giving you a clear view of business impact. And you control the roll out: select which pages to publish optimized versions directly from the dashboard, starting with a short list of priority pages and expanding once you see results.
Agent-Ready Content
BrightEdge Agent Edge gets AI agents to the answer on your page without changing your website. Websites in general are loaded with navigation and technical scripts built for how people read, and that same structure slows an AI agent down before it ever reaches the answer you wrote. BrightEdge research found that 97 percent of the top 20,000 websites are not built to give agents a quick answer. Agent Edge steps in at the same delivery point that already sends your website to every visitor, known as the content delivery network or CDN. For verified AI agents only, it swaps in a clean, lightweight version built from far fewer tokens. Tokens are the small chunks of text an AI system reads and processes. No changes to your website code or CMS templates are required, and human visitors and Googlebot continue to receive the original page unchanged.
Complete AI Responses
Agents look for the answer to the question they were asked, and they try to anticipate what will be asked next. Agent Edge uses BrightEdge intelligence to anticipate those follow-on questions. It adds relevant answers built from content your team has already published, in the fewest possible requests. There is no new content to write and no separate program to run. Answering completely in one pass increases the chance an AI agent uses your content when it creates its recommendations.
Connected Strategy
Agent Edge runs on the same BrightEdge platform that also covers organic search and AI Search performance. Everything you do for AI agents, from setup to monitoring to reporting, lives on that one platform instead of a separate tool. When agent activity shifts, you can see your agent traffic, your content performance and your overall search strategy show up in one connected view.
Ready to Get Recommended
by AI Agents?
AI agent optimization is the practice of structuring website content so AI agents such as ChatGPT, Claude, and Perplexity can read, understand and cite it accurately. BrightEdge Agent Edge does this by detecting verified AI agent traffic and serving those AI agents a lightweight version of a page, while every other visitor continues to see the original page unchanged. That lightweight version strips out navigation, scripts and other elements agents do not need, minimizing tokens, the small chunks of text an AI system reads and processes. Fewer tokens mean agents can read your pages faster and at lower cost, making your content easier for them to cite.
Optimizing for AI agents does not change what human visitors or Google see. BrightEdge Agent Edge routes only verified AI agent requests to the optimized version of a page, while every other visitor, including Googlebot, continues to receive the original, unchanged page.
Most teams set up BrightEdge Agent Edge without any code changes to their website. Setup requires access to their content delivery network, or CDN, and two 30-minute sessions with their IT team to configure that access. Once your CDN access is configured, BrightEdge Agent Edge can be turned on in minutes, starting with a priority list of pages that already receive AI agent traffic.
AI agent traffic can be measured by detecting verified agent requests at the content delivery network, or CDN, and tracking them separately from human traffic. BrightEdge Agent Edge provides a dashboard that shows which agents are visiting, which pages they request and how that traffic changes before and after a page is optimized.
Agent Edge helps enterprises optimize the point of interaction with AI agents across search, AI answers and agent-led discovery
SAN MATEO, Calif.—August 26, 2026—BrightEdge, the global leader in AI-powered organic search, today announced its new search solution Agent Edge, designed to help brands understand, optimize and act on the rapidly growing interactions between AI agents and enterprise websites.
AI agents are evolving from systems that simply retrieve information into increasingly capable digital assistants that research products, compare options, check availability and take multiple steps on behalf of people. BrightEdge believes these interactions are creating a “Second Web” — a new layer of access that creates a second audience for enterprise content: AI agents acting on a person’s behalf, often before that person ever visits the brand’s website, not a replacement for the web built for people or search engines.
The Second Web and The Point of Interaction
At the center of the Second Web is a new point of interaction between the brand and the agent. As AI systems gain more autonomy, the quality of that interaction can influence what information an agent finds, how efficiently it understands a brand and whether that brand ultimately becomes part of an AI-generated recommendation.
Yet most enterprises are not prepared for this shift. BrightEdge research found that 97% of the top 20,000 websites are not currently set up to serve AI agents in optimized formats such as Markdown, which can significantly reduce the amount of content an agent needs to process.
In separate BrightEdge research analyzing ChatGPT recommendations across nine industry categories from June 2025 to June 2026, ChatGPT was found to be researching more but recommending less: the number of sources it cited per answer nearly doubled (+89%), while the number of brands it recommended fell by around 25%. E-commerce saw the steepest decline, with recommendations dropping from roughly six brands to four, while B2B tech held steady at around four.
More than half (61%) of the sites analyzed have implemented schema, an important foundation for helping machines understand website content. But schema alone does not create an agent-optimized experience. Only a small fraction of sites have taken the additional step of serving content specifically designed for AI agent consumption, with early adopters including WordPress, Cloudflare, Linktree and Zendesk.
“AI agents are getting dramatically better at doing the work people once had to do themselves,” said Jim Yu, CEO and co-founder of BrightEdge. “They are researching, comparing and taking action on our behalf, making the point of interaction between an agent and a brand increasingly critical. The companies that can understand and optimize those interactions will have a significant advantage as this second audience grows.”
Giving AI Agents a Faster Path to the Right Information
Traditional websites were built for people and browsers, often requiring AI agents to process complex HTML, JavaScript, navigation and other elements that add little value to machine research. Agent Edge recognizes agent requests at the CDN layer and serves content in a format designed for how agents process information — minimizing wasted tokens and maximizing the value of the information delivered in each request.
Agent Edge then uses BrightEdge intelligence to anticipate the questions an agent is likely to ask next and enrich the content with relevant, data-informed answers — building on the content brands have already invested in rather than requiring them to create a separate experience from scratch. Instead of forcing the agent to return repeatedly to fill in missing context, Agent Edge gives it a more complete response up front, reducing unnecessary processing and minimizing the number of requests needed to complete its research.
In an early deployment with Arm, the semiconductor and software design company behind billions of devices, BrightEdge reduced the payload served to AI agents by approximately 90%. Within weeks, verified AI agent traffic to optimized content more than doubled and, on some pages, tripled.
Another early Agent Edge customer, Bloomfire, saw a 30% lift in direct AI referral traffic following implementation, demonstrating how optimizing the agent experience can translate into measurable discovery and traffic gains.
“We’ve spent years thinking about how people move through our content. Now we also have to think about how agents move through it,” said Brian Zander, VP of Marketing for Bloomfire and early user of Agent Edge. “Agent Edge helps us understand where those interactions are happening, what information agents need, and how to make that experience more efficient without rebuilding the site around a new channel.”
Agent Edge puts the content brands have already invested in to work for AI agents without requiring a separate content operation. Like offering a translated menu to a traveler, it presents the same trusted information in a format designed for how agents consume it.
From there, Agent Edge creates a continuous loop that helps brands:
Together, these capabilities give brands a way to actively shape performance across the Second Web while connecting it back to the search and AI investments already driving visibility today.
Building for a Web That Is Still Taking Shape
Agent Edge extends BrightEdge’s Optimize Once, Win Everywhere vision by bringing agent-led discovery into the same platform brands already use to optimize traditional search and AI answers.
Because these capabilities operate on one platform, the same content, data and intelligence can inform every surface — helping brands avoid fragmented strategies as search and agent behavior continue to evolve.
“The Second Web will not be built all at once,” Yu added. “It will emerge interaction by interaction as agents take on more responsibility for researching and acting on behalf of people. Our goal is to give brands one connected platform to understand and optimize those interactions as they evolve.”
Learn more about Agent Edge and how BrightEdge is helping brands optimize across search, AI answers and the Second Web at https://www.brightedge.com/agent-edge
About BrightEdge
BrightEdge is the global leader in Enterprise SEO and AI-powered content performance. For more than 18 years, BrightEdge has helped thousands of brands and digital marketers, including 57% of the Fortune 500, transform online opportunities into measurable business results. Its industry-first platform integrates the most comprehensive dataset in search, combining insights from traditional SEO, digital media, social, and content with cutting-edge generative AI capabilities, including its deep learning engine DataMind and AI Catalyst platform. Trusted by enterprises, mid-market companies, and leading digital agencies, BrightEdge continues to set the standard for innovation in search and AI, enabling brands to win by becoming an integral part of the digital experience.
CONTACT: press@brightedge.com
AI Overviews expanded into finance in April. Ticker lookups became a larger share of the AI answers people saw. Yahoo Finance's estimated traffic roughly doubled in a single month. By July the most-mentioned brand in the category had changed.
Google is where this story happens. Google AI Overviews carries the volume in finance. Its expansion into the category is what moved the query mix, and it is where the mention lead changed hands. The ChatGPT numbers run the same direction at greater magnitude, from a much smaller base.
The mechanism underneath is classic organic search. Yahoo's ticker pages hold rank one in Google organic results, and those existing rankings aligned closely with the AI citation gains that followed. AI answers are where the shift became visible. Google organic is where the position was already held.
Estimated traffic for Yahoo Finance held close to its July 2025 level for ten months, then roughly doubled between May and June 2026 and settled there. Gradual SEO gains do not produce a single-month step of that size.
| Month | Indexed estimated traffic |
| July 2025 | 100 |
| January 2026 | 104 |
| April 2026 | 101 |
| May 2026 | 99 |
| June 2026 | 221 |
| July 2026 | 214 |
Source: BrightEdge Expert Services. Estimated traffic, monthly, United States, indexed to July 2025 = 100.
BrightEdge's Generative Parser had already recorded a step up in AI Overview presence. In April, Finance moved from one of the industries least touched by AI Overviews to solidly mid-pack, and it held there. The category got AI answers it did not have before.
AI Hyper Cube shows what people were actually asking. In January the highest-volume prompt citing Yahoo Finance was a policy question about tariffs and the stock market. By July the top ten were all ticker and index lookups, entered as queries like "pltr stock", "spy stock", "tesla stock" and "indexdjx: .dji", the Dow Jones index lookup.
| Measure | Growth as a multiple of January 2026 |
| Search volume behind those prompts | 16x |
| Number of distinct prompts | 1.1x |
Source: BrightEdge AI Hyper Cube. Google AI Overview citations of finance.yahoo.com, United States.A narrow prompt set carried far more demand.
Data Cube X shows why Yahoo was positioned to capture the shift. Across Yahoo's ranking keyword footprint, the top-ranking URL is overwhelmingly a dedicated stock quote page: /quote/NVDA/, /quote/TSLA/, one per ticker. No tracked competitor displaces Yahoo at rank one in either snapshot. The asset that won was a page template.
This export is anchored on Yahoo's own ranking keyword set, so competitor absence at rank one reflects Yahoo's dominance within that set. It is not a claim about the whole finance search market.
In January, Forbes led mention share on Google AI Overviews, narrowly ahead of Yahoo Finance. By July that had reversed, with Yahoo at roughly half of all tracked mentions and Forbes down by around a third of its share.
ChatGPT ran a louder version of the same reversal. Forbes gave up more than half its mention share between February and July. MarketWatch rose from almost nothing to roughly a quarter, the largest share movement in this analysis. Google AI Overviews remains the larger surface by volume, so the Google reversal is the one that moves the most business.
The full mention-share movement for all seven tracked finance brands, on both engines, is in the downloadable edition.
The two engines rewarded different things here. In this finance dataset, Google AI Overviews more closely rewarded pages that directly answered the query, which is why a ranking ticker page converted into a citation. ChatGPT frequently named brands as reference points even when it did not cite them.
Forbes's overlap with Yahoo's keyword footprint fell by roughly three quarters between January and July. Forbes did not lose a ranking fight for ticker pages. It was rarely present on them to begin with.
The keyword-overlap figures behind this, for Forbes and for every tracked competitor, are in the downloadable edition.
Because this comparison is anchored on Yahoo's footprint, part of the decline reflects Yahoo's footprint expanding into ticker terms rather than Forbes retreating. Both readings support the same conclusion. The two brands publish different page types, and only one matched the new question.
Some brands are increasingly named by AI engines without being linked to as a source. MarketWatch is named in roughly a quarter of tracked ChatGPT responses in finance. It appears in none of the top cited domains, in either snapshot. Nasdaq and NYSE show the same shape on Google AI Overviews.
The mention-against-citation comparison for MarketWatch, Nasdaq and NYSE is charted in the downloadable edition.
Mention share and citation share are separate assets. They are diverging, and a brand can gain one while losing the other. If you track AI visibility on mentions alone, this is the movement you will miss. AI Catalyst reports both at the prompt level.
1. Check the question before you check the competitor. This shift was demand-driven, not brand-driven. Before concluding that a rival outmarketed you, test whether the questions AI answers in your category moved first.
2. Audit your page format coverage. Establish which page type Google ranks and AI engines cite in your category, then check whether you have one for every entity a buyer might name. Strong organic visibility and the right page format can create an advantage when AI answer demand shifts.
3. Track mentions and citations separately. Rising mentions without rising citations is a distinct problem, and it needs a distinct fix.
4. Plan per engine, and start with Google. Google AI Overviews carries the volume in finance, so that is where scale lives. Google Finance holds roughly 12% of Google AI Overview mentions and nothing on ChatGPT. Visibility is highly platform-specific.
What are AI Overviews?
AI Overviews are the AI-generated answers Google places above traditional organic results. They cite source pages, and the pages they cite are usually pages that already rank well in Google organic search for the same query.
Why did AI Overviews change which brands appear in finance?
AI Overview coverage in finance expanded in April 2026. As coverage grew, the mix of questions receiving AI answers shifted toward real-time ticker lookups. Brands holding a dedicated page for each ticker were positioned for that mix. Brands publishing commentary and list formats were not.
How do you track brand mentions in AI search?
Track mentions and citations as two separate measures. A mention is the engine naming your brand in an answer. A citation is the engine linking to your page as a source. BrightEdge AI Catalyst reports both at the prompt level, and AI Hyper Cube reports share across Google AI Overviews and ChatGPT.
Do Google AI Overviews and ChatGPT cite the same sources?
Not reliably. In this finance dataset, Google AI Overviews more closely rewarded pages that directly answered the query, while ChatGPT frequently named brands as reference points without citing them.
This analysis draws on four BrightEdge sources: Expert Services visibility data for estimated traffic, AI Hyper Cube for prompt-level citations and competitor mention share, Data Cube X for organic rank comparison, and AI Catalyst to corroborate the mention-versus-citation divergence. Each source measures a different keyword universe, so every figure is shown as growth against its own baseline rather than compared across sources. Full methodology, including the anchoring approach and the exclusions applied, is set out in the downloadable edition.