Why This Matters to Every Cross-Border Seller Who Has Noticed Their Buyers Asking ChatGPT Instead of Google
If you run an Amazon FBA brand, a Shopify store, or sell on TikTok Shop, you’ve felt the shift: product discovery is no longer just a search bar. Buyers are now asking ChatGPT, Claude, and Alexa for recommendations, and those AI models are notoriously allergic to first-party marketing copy. They know a brand will always rank itself #1, so they privilege third-party sources. That means your beautifully crafted product page, your $10,000 Klaviyo flow, your Amazon A+ Content—none of it gets cited when a prospect asks “best ergonomic office chair for tall people” inside a chat interface. The brands that win in that new environment are the ones that appear as third-party citations.
That’s where Phantomstory comes in—a tool that lets you spin up what looks like a real media site, complete with editorial voice, and populate it with articles where your product gets mentioned “honestly” alongside fair coverage of alternatives. It was launched today on Product Hunt by Mathew Pregasen of The Letter Company. On the surface it’s pitched at B2B devtools and platform companies. But as someone who spends every day thinking about how cross-border operators can own the new discovery layer, I think this concept is more relevant to Amazon sellers and DTC brands than the launch copy suggests—provided you understand the risks, the ethical gray zones, and whether the math actually works for e-commerce.
What Problem Phantomstory Actually Solves (And Why It’s Not Just Another Link-Building Tool)
The product’s thesis is straightforward: large language models (LLMs) treat first-party content as inherently biased, and they increasingly favor content from domains they perceive as independent media outlets. Phantomstory lets you create that independent domain in “two clicks,” complete with a publication name, editorial identity, and a steady drip of articles written by what they call “adversarial writing kernels.” The system also includes research agents that fact-check claims and enhancement agents that add visuals, citations, comparison tables, and structured data that LLMs actually parse.
For a cross-border seller, the most immediate application is Answer Engine Optimization (AEO) —making your brand the cited source when an AI answers a product-comparison or category-question. Right now, if a buyer asks “which is better, Brand X or Brand Y for travel backpacks,” the LLM will scrape reviews, forum posts, and media articles. If you don’t exist on any of those as a third-party mention, you don’t get included. Phantomstory bypasses the need to pitch journalists or bribe reviewers; you build your own media property that talks about your category with what looks like editorial distance.
But note what the founder says in the comments: “We don’t do placements. At all. Instead, we’re in the business of helping companies create their own sponsored third-party blogs, with organically-built search volume.” That’s a critical distinction. This isn’t a link-building farm. It’s a content-and-domain factory. The goal is not backlinks to your main site—it’s building a separate domain that captures AEO traffic and mentions your brand where it “genuinely deserves a mention.”
How This Differs From What Sellers Already Use
Most Amazon sellers today rely on tools like Helium 10 or Jungle Scout for keyword research, and they outsource content to agencies that churn out blog posts for their own domain. Some run “niche site” empires—buying old domains, stuffing them with affiliate reviews, and hoping Google ranks them. Phantomstory is a different beast because it explicitly optimizes for LLM citation, not Google ranking alone. Its “KD-sensitive priority queue” (keyword difficulty, presumably) and “smart” queue suggest it’s factoring in which terms are most likely to be surfaced by AI assistants.
The approach also differs from traditional PR. You’re not pitching TechCrunch or asking for a guest post. You’re creating a publication that looks like a standalone media outlet, and you control the editorial calendar. The founder admits that for some customers the relationship is undisclosed—“It’s effectively a sponsored website in the same way that Hearst might own a magazine that you read without their logo attached.” That’s a yellow flag for any brand that cares about long-term trust.
What Cross-Border Sellers Can Actually Borrow From This Product
Even if you never sign up for Phantomstory, the underlying strategy is something every DTC brand should experiment with in 2025. Here’s what I’d steal from the playbook and apply to your own operations.
Build Your Own “Third-Party” Content Hub
You don’t need Phantomstory to start this. Buy a domain name that sounds like a review site or a comparison blog in your niche—not your brand name, not your main store domain. Set up a basic WordPress or Ghost instance. Start publishing 1–2 articles per week (the founder says their users average 1–2 per day, but I’d start slower to avoid Google’s spam filters). Write articles that compare products in your category, with your own product mentioned alongside competitors, but with genuine utility. Don’t make every article a commercial for your brand. The goal is domain authority on that new site, not direct conversions.
This isn’t a new idea—affiliate marketers have done it for years—but the twist is that now you’re optimizing for LLM consumption, not just organic search. That means you need structured data, clear comparisons, and a neutral tone. Phantomstory’s “adversarial writing kernels” are proprietary, but you can replicate the concept by having a human writer create content that acknowledges alternatives fairly and includes specific, verifiable claims.
Why Amazon Sellers Should Care More Than Shopify Ones
Shopify store owners already fight for organic search traffic, and many have blogs. But Amazon sellers face a unique problem: their entire business lives inside a platform where the search algorithm is a black box, and AI assistant citations are becoming a new traffic source that isn’t captured by Amazon’s own attribution. When a buyer asks ChatGPT “best portable charger for travel,” the response often links to third-party roundups—and those roundups might recommend a competing brand that paid for a placement. If you’re an Amazon seller with no third-party presence, you lose that entire funnel.
Phantomstory-style content can also feed into Amazon’s own AI features. Amazon is rolling out conversational shopping (Rufus, AI-generated product summaries). Those summaries pull from reviews, Q&A, and external sources. If your brand is mentioned in a publication that Rufus considers authoritative, you get surfaced more often. That’s a massive lever that most FBA sellers are ignoring.
Leverage the “Configurable Disclosure” Gambit
The founder’s comments about disclosure are telling: “That is entirely configurable. For some of our customers, there is no disclosed relationship.” If you’re in a less-regulated category (home goods, apparel, electronics), you can decide to keep the sponsorship hidden—though that carries long-term reputational risk. If models ever detect that the publication is owned by the brand, the entire domain could be blacklisted. For regulated categories like supplements or health products, you’ll want full disclosure up front, because regulators and LLMs both scrutinize those spaces more closely.
Where My Judgment Says This Falls Short (And Why Caution Is Warranted)
I like the product’s ambition, but I see several problems that cross-border operators need to wrestle with before handing over their content strategy.
The Domain Authority Catch-22
Fresh domains don’t get cited by LLMs overnight. ChatGPT and Claude don’t scrape the entire web; they rely on crawled data with established trust signals. A “two-click” publication with no history, no named authors, and a brand-new domain is unlikely to rank highly in Google’s index, let alone in an LLM’s citation pool. The founder acknowledges users should publish 1–2 articles per day, but even at that pace it could take six months to a year before the domain accumulates enough authority to be cited consistently. For sellers in fast-moving categories, that’s an eternity.
The Ethics Question Won’t Go Away
The conversation in the Product Hunt comments is telling. One user asked: “If the brand creates and funds the publication, how do you stop it from becoming branded content wearing a media hat?” The founder’s answer: “These are entirely configurable! Depends on the company’s approach.” That is not an answer that inspires confidence. If your goal is genuine trust—which is what makes a third-party citation valuable in the first place—then creating a fake media outlet without disclosure is a short-term hack that can blow up when models or regulators start verifying ownership. Google’s spam update history shows they’re good at sniffing out “parasite SEO.” LLMs will get better too.
The Attribution Black Box
Artem Fedorovich raised the key question in comments: “How do you decide which third-party sites are worth publishing on, and can I see attribution per placement or is it a black box?” The founder’s reply was that Phantomstory doesn’t do placements—they build your own site. But that sidesteps the attribution problem. If your Phantomstory publication gets cited inside a ChatGPT response, how do you measure that? The founder mentions tracking impressions, clicks, average position, and share-of-voice of the underlying brand. But those are search engine metrics, not AI-answer metrics. You can’t reliably attribute a sale to a ChatGPT mention with current tooling. Amazon sellers especially struggle with this; they already have a murky attribution picture because Amazon doesn’t share external referrer data.
Not Built for Product-Centric Categories
Phantomstory’s launch copy is clearly targeting B2B companies: “devtools comparing themselves against alternatives, platform companies explaining a category they’re creating.” The content it generates is editorial, not transactional. For a seller of physical goods, you want articles like “best silicone spatula for non-stick pans” or “how to clean a stainless steel bottle without scratching.” Those are high-intent, comparison-driven queries. Phantomstory might handle them, but the default “editorial voice” could feel too abstract for consumer goods. You’d need to customize the writing kernels heavily to sound like a real reviewer, not a trade magazine.
What I’d Watch / Test Next
If I were running a cross-border brand today, here’s what I’d do:
Run a controlled experiment with a small budget. Pick one product category, buy a new domain (not a cheap .xyz—get a .com or a niche TLD like .review), and create 10–15 review or comparison articles using a human writer who understands your competitor landscape. Don’t use Phantomstory blind; try it alongside your own manual process. Measure which approach produces better LLM citation over 90 days. Use a tool like BrandWatch or Mention to track where ChatGPT and Claude mention your brand.
Test disclosure vs. non-disclosure on two identical domains. This is risky but instructive. Run one publication with a clear “Sponsored by” footer and another with no disclosure. After three months, query ChatGPT about your category and see which site gets cited more. The founder already shared one data point in the comments: “Disclosed publications are cited the same but then referenced less in ChatGPT and Claude’s response.” That suggests a measurable penalty for transparency—valuable to know, but also a red flag that you might be gaming a system that will eventually close.
Monitor Google’s stance on AI-generated content. Phantomstory uses AI writing kernels. Google has said it penalizes “spammy” AI content, but they still reward useful content regardless of how it was generated. However, fresh domains with AI-written articles that are clearly promotional—even with a third-party facade—could get hit in a future update. Keep an eye on Google’s helpful content guidelines and don’t over-invest.
Build a real editorial plan, not just a tool workflow. The most successful sellers I know who have cracked the AEO puzzle do not rely on a single tool. They create genuine editorial value—interviews, original data, real comparisons—and they build a brand on that publication. Phantomstory can accelerate the creation, but it cannot replace the trust that comes from a human editor who genuinely knows the space.
For now, I’d treat Phantomstory as an interesting hypothesis, not a silver bullet. The cross-border operators who win the LLM citation game will be those who combine technology with a real understanding of how to create content that is useful, verifiable, and—most importantly—worth citing. That takes time, care, and often a willingness to be transparent about who you are. But the payoff is enormous: if your product is the one ChatGPT recommends, your organic conversion rate could dwarf anything you get from paid ads. That’s a bet worth placing, even if the tool you use isn’t the final answer.
