The pitch deck is the new cold email — and most cross-border sellers are still sending the same generic PDF
If you sell on Amazon, run a Shopify store, or manage a TikTok Shop catalog, you already know that the hardest part of growth is not the product. It is getting someone else to talk about the product. UGC creators, affiliate partners, agency clients, wholesale buyers — every one of those relationships starts with a pitch, and almost every pitch I see from cross-border operators is a generic deck with a logo swap. That is why Oriane’s tenth “10 tools in 10 weeks” release caught my attention. Lead Sparker is a thin app built on Oriane’s video intelligence API, and it automates the exact research work that separates a pitch that gets a reply from one that gets archived. For sellers who live and die on earned reach, that distinction is worth real money.
What problem Lead Sparker actually solves
The maker, Julien Rosilio, frames it plainly: he pitches brands and agencies every week, and “the pitch that works is never a generic deck.” It is their brand — the creators already talking about them, the mentions they never saw, what competitors are doing, the trends around them. Pulling that together by hand took him hours per lead. Lead Sparker does the digging for you.
The workflow is almost insultingly simple. Paste the lead’s URL. The Oriane API watches and listens across Instagram and TikTok — untagged appearances caught by visual and transcript search, their feed versus their real earned reach, competitor playbooks, market trends. You get a deck in their brand colors, you sign it with your own note, your name, and your own CTA button, and you download a ready-to-send PDF or email it directly.
That is the whole product. It is a thin UI on top of an API, vibe coded on Lovable, and Julien is upfront about it: “the whole thing is a thin app on the Oriane API.” The API does the seeing and hearing inside videos. The app is just UI on top.
For a cross-border seller, the relevant question is not “is this a clever tool.” It is “does this solve a problem I already pay someone to solve badly.” And the answer, in my experience, is yes — because the alternative is a VA manually scrolling TikTok for three hours and coming back with a screenshot of a tagged post that the brand already saw.
Why Amazon sellers should care more than Shopify ones
Here is a counterintuitive take. The sellers who will get the most out of Lead Sparker are not DTC Shopify operators. They are Amazon FBA brand owners and TikTok Shop sellers.
Why? Because Amazon sellers live in a world where external traffic and off-Amazon brand signals increasingly matter to ranking and to Brand Registry enforcement. If you can walk into a pitch with a slide showing five untagged creator videos that drove 216K views your brand never saw, you are not just pitching a partnership. You are pitching a defensive asset. You are showing a creator or an agency that you understand your own earned reach better than they do.
Shopify operators, by contrast, tend to already have a creator pipeline — they live in Klaviyo flows and influencer dashboards. The marginal value of a better pitch deck is real but smaller. The Amazon seller who has never systematically tracked untagged mentions is the one who will feel the difference immediately.
Where the math breaks
Let me be the skeptic for a second. The Liquid Death example is compelling: a real run found 5 creators who put the brand on camera without tagging it, and 216K views the brand never saw. That slide starts conversations. Julien says so himself.
But 5 creators and 216K views is a single anecdote. The CTO, Thibaut Hadjean, makes the technical case well: finding brand mentions is easy when someone types the name. The whole industry does that already. What is hard is the video where “a can sits on the counter for four seconds and nobody says or writes anything. No tag, no hashtag, no caption.” To every social listening tool on the market, that video does not exist. And those are often the most valuable ones, because nobody paid for them.
Catching those means recognizing the product on screen, across lighting, angles, partial occlusion, and doing it at a volume where finding five instances is statistically meaningful rather than luck. That is the part Oriane says they have spent the most engineering time on, and it is why earned reach numbers from Oriane usually come back higher than what a brand thinks it has.
I believe the technical claim. I have seen enough computer vision demos to know that untagged product detection is a genuinely hard problem. But “statistically meaningful” is doing a lot of work in that sentence. If your product is a commodity — a phone case, a generic supplement, a white-label gadget — visual detection will either miss everything or flag everything. The tool is only as good as the visual distinctiveness of your product and the volume of video in your category. For a niche brand with a recognizable package, this is gold. For a seller pushing the same 10,000mAh power bank as 400 other sellers, it is noise.
How it differs from existing options
The obvious comparison is Brand24, Mention, or Sprout Social. Those tools are built around text listening — keywords, hashtags, brand names, misspellings. They are excellent at what they do, and for many sellers they are the right first purchase. But they share the same blind spot Thibaut names: if nobody types the brand name, the video does not exist to them.
The second comparison is Helium 10 or Jungle Scout on the Amazon side. Those tools tell you about search volume, competitor sales estimates, keyword rankings. They do not tell you who is holding your product on camera on TikTok. Different job entirely.
The third comparison is the agency route. If you have ever hired a UGC agency or a social listening freelancer, you know the deliverable is often a spreadsheet and a deck that looks exactly like what Lead Sparker generates — except it took two weeks and cost four figures. That is the real incumbent here. Not Brand24. Not Sprout. The manual process.
What makes Lead Sparker interesting is not that it is better than Brand24 at text listening. It is that it attacks a category that text listening tools structurally cannot address, and it packages the output as a sales asset rather than a dashboard. That is a meaningful difference. Dashboards are for monitoring. Decks are for closing.
The “thin app” strategy is the real lesson
The most underrated detail in the entire launch is that Lead Sparker is a thin app on an API, built in ten weeks as the tenth of ten free tools. Julien’s cofounder Yuri committed to building 10 free products in 10 weeks on the Oriane video intelligence API. Julien built the first nine. Tool 10 is his. It comes straight out of his own workflow.
That is a go-to-market strategy, not a product. Oriane is not really selling Lead Sparker. Oriane is selling access to the API underneath it. Julien says it explicitly: “If you want to build your own tools on it, happy to get you access.” Thibaut says the untagged detection is exposed directly.
For cross-border sellers, this is the part worth stealing. The most valuable asset you own is not your storefront or your listing. It is your proprietary data about your category — which creators are actually driving demand, which competitors are getting organic lift, which trends are forming before they hit the mainstream. If you can expose that data as an internal API, you can build ten small tools on top of it instead of buying ten SaaS subscriptions. That is how you compound an advantage instead of renting one.
What cross-border sellers can borrow from this
Three things, in order of practicality.
First, the pitch structure. Even if you never touch Oriane, steal the deck logic. Lead with untagged earned reach. Show the creators who talked about the brand without being asked. Show the competitor playbook. Show the trend. Then ask for the meeting. That sequence works because it proves you did work nobody else did. It is the same reason a good Amazon PPC audit beats a generic agency proposal.
Second, the “10 tools in 10 weeks” cadence. You do not need to ship ten products. You need to ship one internal tool every quarter that removes a manual step from your team’s workflow. The compounding effect over two years is enormous. Most sellers never do this because they are too busy fighting fires in Seller Central. That is exactly why the ones who do it win.
Third, the API-first mindset. If you are paying for Klaviyo, Gorgias, and a dozen other tools, ask yourself which one of them is actually just a UI on top of data you could own. The answer is usually more than one.
Where my judgment says it falls short
I have three concerns, and I want to be direct about them.
The first is the launch-day experience. Gal Dayan tried pasting a couple of URLs into it and it kept resetting to a rotating example domain instead of taking his input, so he could not actually see a deck come out the other end. He notes it might just be a launch-day traffic thing. Maybe. But for a tool whose entire value proposition is “paste the lead’s URL,” an input field that resets is a fatal first impression. If you are going to test this, test it on a Tuesday morning, not during a Product Hunt spike.
The second is the detection depth question Gal raises: is it just brand name variants and misspellings, or does it also catch people talking about the product without naming the brand at all — describing what it does instead of what it is called? Thibaut’s answer implies the latter, but the launch copy does not give a clear answer. For a seller with a descriptive product category — “the pink drink,” “the sleep gummy” — this distinction is the whole ballgame. If Oriane only catches visual matches, you are fine. If it also catches semantic descriptions, you have something much more powerful. Not disclosed.
The third is the pricing model. The launch says the tool is free, no signup, as part of the 10-tools challenge. That is a customer acquisition strategy, not a business model. At some point Oriane will need to charge for API access, and the sellers who built workflows on top of it will face a repricing risk. That is not a reason to avoid it. It is a reason to treat it as a prototype, not infrastructure.
The real risk: earned reach is not the same as earned revenue
One more thing. The Liquid Death slide shows 216K views the brand never saw. That is a great conversation starter. But views are not revenue. A creator holding your product on camera for four seconds is not the same as a creator driving conversions. The tool measures visibility, not incrementality. If you walk into a pitch with a deck full of earned reach numbers and the brand asks “what did those views actually sell,” you need an answer. Oriane does not provide one. Your Triple Whale or Northbeam attribution stack does. Use both.
What I’d watch / test next
This week, do three things.
First, go to the Oriane Product Hunt page and run one real lead through Lead Sparker. Not a fake test. A real brand you actually want to pitch. See what comes back. If the untagged detection surfaces even one creator you did not know about, the tool has paid for itself in research time.
Second, audit your own earned reach manually. Pick five creators in your category who have posted about a competitor. Check whether they tagged the competitor. If they did not, you have just found the exact gap Oriane is selling into — and you can decide whether to buy the tool or build the process yourself.
Third, steal the deck structure regardless of whether you use the tool. Build one slide for your next pitch that shows untagged earned reach in your category. It does not need to be perfect. It needs to prove you did work nobody else did. That is the actual product. The API is just the fastest way to get there.
