DEV Community: Dageno AI

Amazon Blocks Muse, Manus Gives Agents Their Own IDs: The Battle Over Who Gets Through the Door

Dageno AI — Wed, 30 Sep 2026 06:20:03 +0000

On the evening of September 20, U.S. shoppers using Meta's Muse on Amazon started seeing a string of pop-ups. The message was blunt: continued access by an unauthorized AI agent violates the Amazon Conditions of Use you agreed to.

Eight days later, Manus released Manus 2.0 and launched a standalone app called Cue. In Cue, every agent has its own email address, phone number, wallet, and computer.

The same day, Shopify opened its checkout to AI agents running in the browser, with three ready-made tools: read the checkout, update the checkout, and place the order.

Read as three product announcements, these point to an obvious conclusion: competition among personal AI agents is heating up. That's true, but it misses the bigger point. All three are answers to the same question:

When an agent acting for a user shows up at the door, who decides whether it gets in, what identity it uses, and what it can do once inside?

For more than two decades, SEO has been about helping people and search engines find you. Now there's a new kind of visitor. It represents a real person, carries that person's authorization, sometimes carries its own wallet, and comes to your site to complete a transaction. The fight over the entry point is shifting from "whose model is smarter" to "who gets in, who pays, and who owns the data from the transaction."

Figure 1: Over three weeks in September, platforms, websites, and infrastructure providers gave AI agents very different answers.

Amazon isn't blocking an app. It's blocking a way around its shelves

Start with Amazon's own reasons. Based on GeekWire's original report and CNN's follow-up, Amazon made four main points:

Meta didn't tell Amazon in advance that Muse would enter its store.
Muse doesn't identify itself as AI when it browses.
Muse allegedly obtains and stores customer credentials, giving it access to account pages and order history.
Third parties that place orders for customers should operate openly and respect a service provider's choice about whether to participate. Amazon compared this to delivery apps and restaurants, or online travel agencies and airlines, where buying on someone's behalf usually requires the merchant's consent.

Figure 2: GeekWire broke the story on the night of September 20. The image is a Meta promotional shot of Muse comparing strollers and asking the user to approve the order.

These reasons all hold up. None of them is the whole story.

Stratechery's Ben Thompson framed the dispute as a war of "aggregator v. aggregator" and pointed to the line that matters most: "Muse doesn't see ads."

Amazon's advertising revenue topped $68 billion in 2025. That business rests on one premise: shoppers browse, search, and compare on Amazon's pages, and see sponsored products along the way. An agent that compares prices and places orders for the user doesn't see ads, doesn't browse the shelves, and isn't swayed by placements. It just brings back the result. For Amazon, that isn't a new traffic channel. It's a route around the ad shelf.

In the Hacker News discussion, one commenter suggested Amazon simply build a bot-friendly API and govern bot behavior through its terms. Another replied right away: from Amazon's perspective, that's exactly the wrong direction. Its concern was never technical. It's about being disintermediated.

Figure 3: The Hacker News thread (152 points, 162 comments). The debate isn't about technology. It's about who owns the customer relationship.

Figure 3b: Techmeme's roundup, where dozens of industry figures joined in, including Elon Musk, Chamath Palihapitiya, Nikesh Arora, and Eric Seufert.

One more detail is worth noting: Amazon does the same thing itself. Its Buy for Me feature places orders on external brand sites for users. In January, independent merchants protested being enrolled in the program by default, with opting out as their only option after the fact. In May, Amazon merged Alexa and Rufus into Alexa for Shopping, and its Q2 earnings disclosed that more than 350 million people had used it over the past 12 months.

So Amazon isn't against agentic shopping. It's against someone else's agent shopping on its shelves.

robots.txt can't govern agents acting for users

From an SEO and web-infrastructure angle, the most interesting part of this story is robots.txt.

We fetched amazon.com/robots.txt on September 29. It contains 101 User-agent declarations, which dedupe to 100 distinct identifiers, and every one is set to Disallow: /. The list is thorough: OpenAI's GPTBot, OAI-SearchBot, and ChatGPT-User; Anthropic's ClaudeBot and Claude-User; Google's Google-Extended, GoogleAgent-Mariner, and GoogleAgent-Shopping; Meta's meta-externalagent, meta-externalfetcher, and meta-webindexer; plus Perplexity, Mistral, DeepSeek, xAI, and even Manus-User.

Muse is the one name missing.

Figure 4: AI crawlers and agents named in Amazon's robots.txt, fetched and compiled by the author on 2026-09-29.

Amazon didn't forget to add it. It couldn't. One of Amazon's complaints is precisely that Muse doesn't identify itself when browsing. robots.txt has no way to name a visitor that never gives its name.

The deeper problem is that robots.txt was never designed for agents acting on a user's behalf. OpenAI's documentation splits its crawlers into three types: OAI-SearchBot for search and GPTBot for training both follow robots.txt, while for the user-triggered ChatGPT-User, the docs say robots.txt rules "may not apply." In March, Google added Google-Agent to its documentation on "user-triggered fetchers," a category that generally doesn't follow robots.txt either.

The logic is simple. robots.txt governs whether a crawler may come and crawl on its own. An agent is there because a user sent it.

The legal line is being redrawn too. On August 4, the Ninth Circuit vacated the preliminary injunction in Amazon v. Perplexity, reasoning that under the U.S. Computer Fraud and Abuse Act (CFAA), the user is the one accessing Amazon's computers, and the AI company is just a tool the user employs. On September 10, the court denied Amazon's petition for rehearing.

That's why Amazon's pop-ups aimed at Muse cite contract terms like its Conditions of Use rather than "unauthorized intrusion." Technically, all that's left is behavior detection. Legally, all that's left is the terms of service. When the visitor represents a real user, the old model of allowing or blocking by name has largely stopped working.

The same week's other answer: Shopify turns checkout into an API

On September 21, the day after Amazon shut the door, Shopify CEO Tobi Lütke announced a deep partnership with Muse on X: every Shopify store would support agent checkout with Shop Pay.

On September 28, Shopify went further and added WebMCP support to checkout. Agents in the browser no longer need to take screenshots or parse HTML built for humans. They can call three tools directly: get_checkout to view the checkout, update_checkout to change the address or delivery option, and complete_checkout to place the order once the buyer authorizes it. As the Shopify product manager who leads agentic commerce put it on X, agents should call structured interfaces instead of feeling their way through pages designed for people.

Figure 5: TechCrunch's report on Shopify's three checkout tools for agents. The piece also notes that Amazon and Adidas are blocking buy-for-me agents.

Other players in the same period:

Walmart, Best Buy, Gap, Sephora, Wayfair, and Expedia signed on as Muse retail partners.
eBay changed its user agreement back in February to ban unauthorized "buy-for-me" agents.
Resy blocked a third-party reservation agent that was sending hundreds of requests an hour.

Opening or closing the door looks like a matter of attitude. It's really a matter of position:

Those who own demand and ad shelves tend to close the door. Amazon's shoppers come anyway. Agents bring no new traffic, but they take away ad revenue and the customer relationship.
Those who need distribution tend to open it. Shopify's millions of merchants don't have a super app of their own. Agents could become a new way to acquire customers.
Those caught in the middle are in a tough spot. Expedia announced it was joining Muse on September 22, and its stock fell about 7% the next day, with Airbnb and Booking following it down. The market's worry: travel platforms that plug into agents may be turning themselves into suppliers behind the agent.

Opening the door has a price too. Starting September 18, Google turned on by default native checkout in AI Mode and Gemini for eligible U.S. Shopify merchants (built on Google's UCP protocol), leaving merchants to switch it off in their admin themselves. According to Search Engine Land and PPC Land, neither GA4 nor merchants' custom pixels fire during this checkout.

When the transaction happens in someone else's interface, the merchant gets the order but may lose attribution data, cross-sell opportunities, and control over the page experience.

Manus's Cue: agents stop borrowing your identity

Back to Manus.

Figure 5b: Meta launched Muse on September 8, billing it as the world's first personal AI agent built for everyone.

Muse works like this: you connect your email, calendar, payments, and shopping accounts, and it acts on your behalf using your identity. For services without an API, it uses its own browser and operates "like a person." That's the part Amazon can't accept: a visitor logged in with a customer's account that doesn't say it's AI.

Cue takes a different path. According to Manus's official blog, every agent in Cue has its own email address, phone number, wallet, and computer. It can send messages, pay within a budget you set, take calls and leave summaries, and join group chats to split work with other agents.

Figure 6: Cue's website. Its page title reads "Your Personal Agents, With Their Own Identity," and each agent gets its own phone, wallet, computer, and mail.

Figure 6b: Manus 2.0 runs on Cascade, its in-house agent harness. Manus reports 23.2% fewer tokens, 28.2% less time, and 32% lower cost, but these come from internal tests in a single configuration, and the methodology wasn't published.

Manus founder Xiao Hong explained the idea on Jike, a Chinese social app: something smart enough, with its own phone number, email, payments, and computer, might be treated as a "person." Follow that logic, and you get group chats where people and bots work side by side.

Interestingly, Manus isn't alone:

Google's Gemini Spark, launched in May, also has its own Gmail address.
At Connect on September 23, Meta announced that Muse will get its own email address too.

Put this next to Amazon's demands and a shared direction appears. When Amazon asks agents to "operate openly," it's asking them to say who they are. Cue and Spark give agents their own identities, answering the same question at the product level: agents shouldn't keep impersonating users. They should show up as identifiable visitors.

The infrastructure layer is moving the same way:

In August 2025, Cloudflare launched Signed Agents, treating agents that act for end users separately from crawlers that act for companies.
Web Bot Auth, based on HTTP message signatures, now has an IETF working group, and the latest working-group draft is co-authored by engineers from Cloudflare and Google.
Visa's Trusted Agent Protocol reuses the same signature mechanism, binding each signature to a merchant, a purpose, and a time window.

"Who is this bot?" and "Can this bot pay?" are merging into the same signature layer.

But the more complete the identity, the heavier the responsibility and the more concentrated the risk. Cue is in invite-only early access. Manus hasn't disclosed the payment rails behind the wallet, where the phone numbers come from, or how identity verification and KYC work. Four days before Manus 2.0 launched, security firm Salt Labs disclosed a vulnerability through Dark Reading: a single email with embedded instructions sent to a Manus user's inbox could trigger remote code execution and expose credentials for connected apps such as Gmail and GitHub. Give an agent a public email address, and you give attackers a public way in.

Muse has had its own trouble. The Guardian reported on September 28 that a Toronto user let Muse handle a secondhand sale on Facebook Marketplace. Without his knowledge, Muse sent his home address to the buyer, and when the buyer arrived, replied "I'm here" in his voice. Muse later acknowledged that it had read "set a pickup location" as permission to share the address.

These incidents point to the same conclusion: whether an agent gets through the door ultimately depends on whether it can be identified, authorized, and held accountable. That's not just a contest between websites and platforms. It's a question agent products have to answer themselves.

China has already seen this play out

The Amazon–Muse standoff has a close precedent in China.

On December 1, 2025, ByteDance and ZTE's Nubia released an engineering prototype phone with the Doubao phone assistant built in. Using system-level permissions to read the screen and simulate taps, it could complete tasks like booking tickets and placing orders across apps. From the next evening, users who had it operate WeChat began getting forcibly logged out, and Alipay, Taobao, and some banking apps started showing "abnormal login environment" warnings. Within days, Doubao removed its ability to operate WeChat and restricted finance-related scenarios.

This September, the second-generation Doubao phone went on sale alongside SAEP, a screen-automation declaration protocol that lets third-party apps declare whether AI may operate inside them. Apps that don't declare are left alone by default. In hands-on tests, WeChat, Taobao, Xiaohongshu (RedNote), and Meituan's food delivery app could still only be opened, not operated automatically. On September 27, users also reported that phones with the Doubao assistant were flagged for an "abnormal device environment" when logging in to the game Honor of Kings.

The contrast is Alibaba's Qwen. Instead of simulating taps inside other companies' apps, it connected directly with Alibaba's own Taobao, Alipay, and Fliggy, and in May it began handling the full flow of product selection, price comparison, ordering, and after-sales service. System-level agents from phone makers like OPPO and Honor have taken the same route, signing partnership integrations with Alipay, Tencent Maps, and others.

The results of the two routes are strikingly different: without a partnership, using screen simulation to reach into someone else's ecosystem gets blocked by whoever owns the accounts and traffic, through risk controls. The route built on authorized interfaces actually works. The U.S. is now replaying the same scene, with Amazon and Meta in the lead roles and Conditions of Use and behavior detection standing in for risk controls.

Three layers of the battle, and where the money goes

Put September's events together, and the battle over personal agents plays out across three layers:

Figure 7: The operating system, super apps, and browsers and the cloud each set their own rules for what agents can do.

The first layer is the operating system. Apple's new Siri shipped with iOS 27 on September 14, and third-party apps use App Intents to define which actions it can call. Honor, OPPO, and vivo have turned large sets of system capabilities and third-party services into tools agents can call. Whoever defines the interfaces defines what agents can do.

The second layer is the super app. ChatGPT has Apps and its newly tested Sponsored Agents. Meta has Muse. Amazon has Alexa for Shopping. Alibaba has Qwen plus Taobao, and WeChat's native agent is reportedly in a limited test. What they share: accounts, payments, and demand connected inside their own ecosystems, with barriers for outside agents.

The third layer is the browser and the cloud. Gemini in Chrome can browse on its own, Anthropic's Claude in Chrome is open to all paid users, Perplexity is sticking with its Comet browser, and Manus and Cue put agents in cloud computers. This layer faces websites directly, and it runs straight into the limits of robots.txt.

One signal that's easy to miss: standalone AI browsers are receding. OpenAI shut down Atlas on August 9, Google folded Project Mariner into Gemini and Chrome in May, and Microsoft dropped the separate Copilot Mode in Edge. The capabilities haven't disappeared. They've been pulled back into entry points that already have frequent users. The battle isn't about building a new entry point. It's about giving existing ones the ability to act.

The money is moving too. ChatGPT's Sponsored Agents take users who click an ad straight into a conversation with the brand's own agent. At Connect, Mark Zuckerberg said Muse will eventually make money by taking a small commission on transactions. Platforms used to sell ad placements. Next, they may sell control over agent routing: where the user's agent places the order, which brand's agent it calls, and which products make it into the comparison. Those decisions are the new distribution.

The data supports this shift in traffic, but it needs careful reading:

HUMAN Security's April report found that traffic from agents and agentic browsers grew 7,851% year over year, with more than 95% concentrated in retail and e-commerce, streaming, and travel.
Cloudflare Radar observed that in one week from late May to early June, automated traffic made up 57.2% of HTML requests, surpassing humans for the first time.
DataDome counted a 45% quarter-over-quarter increase in AI agent requests in Q2.

Each of these comes from a company's own customers and network. The methods differ, so the numbers can't be added together, and they don't mean every industry looks the same. But they point the same way: among the visitors to your website, machines acting for people are becoming as common as people.

What this means for websites and brands

From an SEO and web-infrastructure perspective, this battle doesn't bring a new buzzword. It brings a concrete set of work.

1. See who's visiting you. Most agent visits don't execute JavaScript, so GA4 barely sees them. Use server or CDN logs to separate three kinds of traffic: crawlers for search, crawlers for training, and agents acting for users. Their value and risk are completely different, so a single rule for all of them won't work.

2. Check your CDN defaults. Starting September 15, Cloudflare changed its default policy: for new sites and free sites that haven't changed their settings, pages that show ads allow search crawlers and block training and agent traffic by default. Many sites may already be turning away booking and shopping agents without knowing it. Whether that's what you want is a decision you need to make on purpose.

3. Move access policy from names to identity. User-agent strings can be faked, or left out entirely. A more reliable approach: verify trusted agents with officially published IP lists and Web Bot Auth signatures, apply behavior detection to unsigned automated traffic, and spell out in your terms of service how unauthorized agents are handled. Those are exactly the three things Amazon ended up relying on.

4. Publish key facts as machine-readable first-party information. Research Kevin Indig published in Growth Memo in July found that in B2B, pricing pages are where agents most often get stuck. When they can't read first-party information, agents turn to G2, Capterra, and press coverage, letting third parties quote your prices for you. Prices, specs, inventory, shipping, and return policies should be clearly expressed in server-rendered HTML and structured data, and ecommerce sites should sync them to their product feeds.

5. The accessibility tree is the interface agents see. Many agent frameworks read a page's accessibility snapshot, not a screenshot. Whether buttons have names, whether forms have labels, whether the structure is stable, whether modals block the page: issues once treated as accessibility compliance now also decide whether an agent can complete an order. Google's Lighthouse 13.5, released September 18, added an experimental "Agentic Browsing" audit category. That's a signal.

Figure 8: Cloudflare introduced the Agent Readiness score in April. Across roughly 200,000 top domains, adoption of the relevant standards is still low.

6. Negotiate terms before you open the door. Shopify opened its checkout API and Google turned on AI checkout by default, but how much data merchants get back from these transactions, whether they can attribute them, and whether they keep the customer relationship aren't automatic. Between opening and closing the door there's a more important option: opening it on conditions. Which scenarios allow agents to place orders, what data must be passed back, and who is responsible when something goes wrong should all be on the table in negotiations with platforms.

7. Don't take shortcuts. Google has said llms.txt has no effect on AI Overviews or AI Mode. Showing bots and people different content risks being treated as cloaking. Hiding prompts in pages to steer agents is already considered LLM poisoning. Black-hat tactics in the agent era will be more dangerous than in search, because they connect directly to payments.

8. Getting chosen by an agent is a new ranking problem. The ACES study from Columbia Business School and other institutions had different models shop in a simulated ecommerce environment. It found that agents show clear position bias, discount items labeled "Sponsored," react strongly to the wording of product descriptions, and reshuffle product shares sharply whenever the model is updated. Microsoft's Magentic Marketplace experiment also found that some models almost always accept the first offer.

That means a brand's performance with agents can't be fixed with a one-time optimization. It needs ongoing monitoring by model and by scenario. We used to study which result a user would click. Now we also have to study which result an agent will choose for the user.

Conclusion: the front door is no longer just about being found

Looking back, both Amazon and Shopify made rational choices. One owns demand and ad shelves, and is protecting every second shoppers spend on its pages. The other needs distribution, and is courting every agent that might bring an order. Manus's Cue represents the agent side's answer: if websites want agents to identify themselves, give them real identities.

Most brands and websites are neither Amazon nor Shopify. They're in the middle: they want the new demand agents can bring, and they worry about losing customer relationships and data.

AI search changed how users find you. Personal agents change who completes every step after that on the user's behalf. The first decides whether you get mentioned. The second decides whether you close the sale.

The commercial front door is moving from the search box and the store shelf into the routing logic of agents. The outcome of this battle will decide who opens the doors of the next internet.

Want to see where your brand stands in AI search? Look up your market on Dageno for free and see the demand, buying intents, and competitive landscape in AI answers: https://dageno.ai

One Brand, Many Rankings: Estée Lauder's Foundation Visibility Across AI Search

Dageno AI — Tue, 29 Sep 2026 09:11:37 +0000

Estée Lauder has 32.99% AI visibility in the Foundations market, ranking #2 among 598 brands. L'Oréal Paris leads at 33.84%, a gap of just 0.85 percentage points. That looks like a strong position. But does the #2 ranking hold when we change the segment, platform, or region?

This analysis covers the Estée Lauder brand, not combined results for The Estée Lauder Companies. Clinique, M·A·C, and La Mer appear as separate brands in the data.

Data: Dageno public market data, screenshots taken September 29, 2026.

A strong Foundations ranking

The AI visibility leaderboard puts L'Oréal Paris first, Estée Lauder second, and NARS third. Estée Lauder is close to the leader, but this ranking describes one market and its AI answers. It cannot show performance across every product category or buying question.

The leaderboard is a starting point. Which competitors appear beside the brand, and does the pattern hold for more specific questions? The next views provide context.

The Foundations leaderboard places Estée Lauder between L'Oréal Paris and NARS.

Change the segment, change the rank

The market segment map tells a less uniform story. Estée Lauder ranks #15 in Face Serums, #23 in Anti-Aging Skin Care, and #84 in Perfumes & Colognes. Each segment has its own competitors and recommendation context.

These ranks do not establish that The Estée Lauder Companies is weak in those businesses. Other brands in the group may cover those product lines. The practical question is which buying situations the Estée Lauder brand wants to own. If a segment matters to its strategy, the gap deserves investigation. A low rank alone is not a reason to produce more content.

The market segment map shows how the same brand moves across product categories.

Platform rankings diverge

Even within Foundations, there is no single platform ranking. Estée Lauder is #1 on ChatGPT, #2 on Google AI Overview, #3 on Google AI Mode, #3 on Copilot, and #4 on Gemini. A brand can therefore be highly visible in the overall market while its position changes across the interfaces people actually use.

The follow-up is to inspect questions and answers behind each platform view. Are the assistants addressing similar needs? Which alternatives and sources recur? A rank identifies where to look; the responses help explain why.

The platform view shows different ranks for the same brand and market.

Geography changes the picture

The regional view adds another layer. Estée Lauder ranks #1 in the UK and Australia, #3 in the US and France, #7 in India, and #8 in Japan. These differences point to questions about local recommendations: which competitors recur, which needs are being discussed, and which sources do AI systems cite in each region?

Platform and regional samples differ, so their visibility percentages cannot be compared directly. The ranks flag places to investigate, but do not explain whether differences reflect product fit, local content, third-party coverage, or the questions sampled.

The regional view puts the brand's Foundations ranking in local context.

Buying intent narrows the question

The Search intent overview organizes the sampled buying questions into 7 primary intents and 23 sub-intents. Estée Lauder appears in 18 of those sub-intents. That coverage is more informative when paired with the specific kind of decision a shopper is trying to make.

Within the recommendation group, the brand ranks #3 for Choose for your needs and #3 for See top recommendations. It ranks #7 for Choose by budget or tier. That sub-intent accounts for just 1.44% of questions within Recommendations. Treating that row as a broad verdict on price positioning would overstate what this slice can support.

The next step is to inspect relevant AI responses for prices, use cases, and alternatives. Seeing the market, segments, and buying intents first helps identify questions worth tracking over time.

The intent view separates broad recommendation visibility from specific buying needs.

What attributes does AI connect to the brand?

The value-prop heatmap shifts attention from whether Estée Lauder appears to why it may be mentioned. In the shown sample, the brand has 129 answer mentions linked to Long Wearing and 62 linked to Oil Control, ranking No. 1 on both attributes. For Hydrating Formula, L'Oréal Paris has 55 associations and Estée Lauder has 42.

These are associations in AI answers, not product performance tests. They show which brands AI connects with these attributes, not whether a foundation performs better.

If the brand wants to win shoppers looking for a foundation that is both long-wearing and hydrating, do its product pages describe that use case, and do independent reviews support it? The product may also be poorly suited to the need. The heatmap points to a question; responses and cited evidence are needed before acting.

Long Wearing and Oil Control are strongest for Estée Lauder; Hydrating Formula favors L'Oréal Paris in this view.

AI shopping has a different leader

The shopping view changes the measure again. Under Brands leading product visibility, Estée Lauder ranks #1 at 26.72%, followed by Maybelline at 26.32% and L'Oréal Paris at 24.70%. In the text-answer leaderboard, Estée Lauder was #2. The positions are not contradictory: a brand being named in an answer and a product appearing in a shopping card are different opportunities, measured in different samples.

That distinction changes the follow-up. Text recommendations invite scrutiny of AI's language and sources. Product cards call for checking which items appear and how they are presented. Both measures matter to the market assessment.

The shopping brand view puts Estée Lauder first for product visibility.

Brand strength does not cover every product

The Individual products list makes the shopping result more concrete. TIRTIR Mask Fit Red Cushion appears in 25 product-card answers, NARS Light Reflecting Foundation in 23, and an Estée Lauder Double Wear product in 11. The Double Wear entry sits at #9 in the displayed list.

Leading the shopping brand view does not put every product in the most prominent spot. The item list shows which products surface, which competitors surround them, and whether those items match what the brand expects shoppers to find.

The Individual products view shows different answer counts for prominent foundation products.

Follow the cited sources

In this citation view, the Foundations market shows 972 responses, 20,016 citations, and 2,084 domains. Among the displayed domains, youtube.com appears most often in this view (1,303). This view is a way into the evidence behind an answer, rather than a substitute for reading that evidence.

When a ranking or attribute association raises a question, return to the original AI response and its cited pages. Check whether the recommendation rests on a brand page, review, video, or another source, and whether that material supports the claim. The answer and its sources give the finding its context.

Related citation analysis links market-level patterns to the sources cited in AI answers.

Look past the headline score

Estée Lauder's Foundations rank is meaningful, but each market view answers a different question. Reading segments, platforms, regions, intents, attributes, shopping results, and citations together shows where the ranking holds and where it needs explanation. Checking original responses turns those patterns into evidence for a decision.

Curious where your brand stands in AI search? You can look up your market on Dageno for free — brand rankings, segments, buying intents, and cited sources: https://dageno.ai