Every dollar of margin in cross-border e-commerce is negotiated before it’s earned. The factory quote, the freight quote, the repricing decision, the next campaign bid — each one happens in a conversation where you can’t pause, Google, or ask a friend for a second opinion. We’ve built entire tool stacks for the before and after of those conversations, but almost nothing useful for the during. That’s why Sorinai, the interactive AI notepad from the company of the same name, caught my attention. It’s not another transcription widget. It’s a meeting tool that sits beside you while the call is live, fills in your own templates, and answers private questions like “What should I ask next?” For operators whose best deals were made in a supplier call and whose worst losses were too, that’s not a convenience. It’s leverage.
The Problem Isn’t Note-Taking, It’s Timing
Every AI notetaker I’ve tested — Granola, Jamie, Superpowered — is a playback machine. It records, transcribes, and hands you a summary after the meeting ends. That’s useful for people who rarely make live decisions. But in cross-border trade, the live decision is the whole game. The moment a factory says “USD 4.60 FOB Shenzhen,” you either catch the implication or you don’t. The answer only matters while the supplier is still on the line.
Founder Thomas puts it more honestly in his launch note: “AI has become very good at helping us before and after the things that matter-but almost useless during them.” That’s the core observation here. Sorinai’s response is to be present during the call. It captures your computer’s audio, so no bot has to join the meeting, and it works whether you’re on a video call, a phone call, or sitting across the table from someone. It reads the template you already use — Word, PDF, or plain text — and fills every field as the call progresses. At the same time, you can ask it real-time questions: “What did that acronym mean?” “Is what they just said true?” “What should I ask next?” — the private help you’d never ask out loud.
That last feature is the one that matters for cross-border sellers. On a call with a sourcing agent, you don’t want to reveal that you don’t know what a “surcharge” really covers. You don’t want to expose that you’ve never negotiated incoterms before. You want a quiet whisper in your ear telling you what to probe next. Sorinai is trying to be that whisper.
Why Amazon sellers should care more than Shopify ones
Not every operator needs this equally. Shopify DTC teams live in dashboards: ad spend, conversion rate, AOV, email flows in Klaviyo. Their meetings matter, but the decisions are tested daily by data. Amazon sellers, by contrast, live inside conversations that produce documents: the supplier quote becomes your landed cost, the Amazon Seller Central case log becomes your defense, the FBA inbound plan becomes your purchase order. Template discipline is already survival.
That’s why template-filling is the sleeper feature. A typical product launch meeting for an Amazon FBA product has a predictable skeleton: unit cost, MOQ, packaging spec, QC pass/fail, lead time, payment terms, shipping method. If an AI notepad can read that skeleton and fill it in while the factory is talking, then the post-call work shrinks to review, not reconstruction. Shopify operators have more agency meetings and fewer document-heavy operational calls. They’d benefit, but Amazon and Walmart marketplace sellers — plus private-label sellers who still deal with factories — feel this pain first.
How It Differs From the Existing Notetaker Clique
The incumbents are all trying to replace the note-taker. Sorinai is trying to replace the note itself.
Granola is “the AI notepad for people in back-to-back meetings.” Jamie calls itself “human-quality meeting summaries without a bot.” Superpowered is “the AI copilot for your meetings.” HyNote AI and Speech to Note round out the same idea. All of them produce clean output. But they operate after the fact, and the output is a machine summary detached from your own mental model of the deal.
Sorinai’s real distinctions are four:
- Template-first. It doesn’t generate a generic summary. It fills in the structure you already use. That’s a subtle but huge shift: the output is usable in your existing workflow, not another document to reconcile.
- Live interactivity. You can ask questions during the call while the answer still matters. That changes the unit of value from a summary to a decision aid.
- No bot. It captures local audio, so there’s no visible participant, no “this meeting is being recorded” interruption, no awkward reaction from the supplier.
- Cross-meeting memory. You can query one meeting or all of them at once. For a seller juggling three factories and two freight forwarders, that’s the difference between memory and a database.
The no-bot detail is more important than it sounds in cross-border work. Many sellers tell me their suppliers are wary of visible meeting bots. A bot joining a Zoom call or a Teams call creates friction before a single word is spoken. Sorinai’s approach — listen to the computer’s audio without joining as a participant — removes that friction. But it also raises a question that the Product Hunt comment section already anticipated. One commenter, Arash Rahimi, asked how speaker attribution works on a four- or five-person call if the tool is only reading system audio. That’s exactly the right challenge. On a chaotic call with a factory, a sourcing agency, and a freight forwarder, knowing who said what is where trust is built. If Sorinai can’t attribute cleanly, its template-filling becomes less useful the moment three voices start negotiating over each other.
What Cross-Border Sellers Can Borrow (Even If You Never Install It)
Whether or not Sorinai becomes a permanent part of your stack, its design contains a few lessons worth stealing.
Build a negotiation template before the call. Not a list of questions — a document with blanks. Sorinai reads Word, PDF, or plain text and fills every field from the call. But you don’t need Sorinai to benefit from this habit. Most sellers show up to supplier calls with a mental checklist. A better approach is a form with mandatory fields: target unit price, confirmed MOQ, payment terms, incoterms, lead time, packaging spec, sample deadline, QC inspection date, exclusivity terms, next action. When the call ends, the form is either complete or you know exactly what’s missing. That alone is worth more than most meeting-summary tools.
Keep your notes in your own words. The founder’s best line is this: “Most tools take the meeting away from you and hand back a summary written by a machine. sorinai works the other way around: you write what you notice, in your words, and it tidies and fills in around that. What you thought was important stays important.” For cross-border operators, that’s a philosophy, not a feature. Your own annotations about a factory owner’s hesitation or a freight forwarder’s vague answer are signals no AI generated summary will preserve.
Use private real-time questions as a negotiation coach. “Is what they just said true?” is a brutal question to ask out loud in front of a supplier. But it’s exactly what a good procurement mentor would whisper in your ear. Same with “What should I ask next?” In a negotiation, the biggest mistakes come from the questions you didn’t ask because you didn’t want to look inexperienced. A private AI that can prompt you without exposing you is a genuinely interesting use case.
Treat every meeting as part of a corpus. Sorinai lets you query one meeting or all meetings at once. That’s the habit to copy. If you’re comparing vendor promises over the last quarter, you shouldn’t rely on memory or scattered notes. You should be able to ask: “Which factory promised a six-week lead time?” and get a cross-call answer. You already use Helium 10 to systematize product research. The next gap to systematize is call intelligence.
Where the math breaks
The obvious concern is sustainability. The launch page says Unlimited notes, free. That’s a generous hook, but a business needs revenue. There’s no disclosed paid tier, no disclosed business model. For users, free is great. But it also means the roadmap will be shaped by whatever comes next — venture capital, a subscription pivot, or feature cuts — rather than by your needs as a seller.
The second concern is platform reach. The page says Mac and Windows. My life is on a laptop, but a large share of cross-border seller life happens on a phone: WeChat voice messages, WhatsApp calls, a quick call with a freight agent while you’re at a trade show. If the tool can’t capture phone audio, it’s useless for exactly the conversations that are most chaotic and most important. I’d want to test that before trusting it in the field.
The third concern is privacy. Capturing computer audio means confidential negotiation details are passing through another piece of software. How that audio is stored, whether it leaves the machine, and who has access are all questions the launch page does not answer. For sellers operating under non-disclosure agreements or sourcing in sensitive categories, that’s not a small detail.
My Honest Judgment: Useful, But Not Yet a Trade Workflow
This is Sorinai’s second launch, and the launch presence is modest — 71 followers at the time of the page, sitting around the daily leaderboard at roughly the #20 slot. That’s not a criticism. Early-stage meeting tools often need a few rounds to find product-market fit. But it means you’re not adopting a mature platform; you’re adopting an experiment.
Still, the direction is smarter than the category. Most meeting tools optimize for the transcript. Sorinai optimizes for the decision. The template angle is especially clever for operators because it ties the AI output to a real business document rather than another note-taking artifact. If a future version could fill a PO, a QC checklist, or a supplier scorecard and then export it straight into your operations stack, it would stop being a notepad and become infrastructure.
For now, I’d treat it as a powerful experiment, not a core-stack tool. The absence of mobile capture and the open speaker-attribution question are real gaps. And as far as the launch page discloses, there’s no mention of multilingual capability. If Sorinai could fill an English template from a Mandarin or Cantonese call, it would be the single best sourcing tool I’ve seen in years. Until then, it’s a promising notepad with a few sharp edges. That’s still more interesting than another transcription app.
What I’d Watch / Test Next
The fastest test: take your next supplier call and turn your usual checklist into a Word template with blanks — MOQ, unit price, incoterms, payment terms, lead time, packaging, QC plan, exclusivity, next action. Run Sorinai in the background and see whether it fills those fields while the call is live. During the call, ask it “What should I ask next?” If the answers are generic, ignore them. If one catches a term you’d normally forget, keep the tool in the loop.
Second, test the truth-check on a call where someone is making a questionable claim. Ask “Is what they just said true?” and see whether it catches inconsistencies or repeats the claim back without evaluating it. Third, after two weeks, try querying all meetings at once to compare vendor promises. And fourth, ask the maker directly — Thomas says he reads every message at [email protected] — about mobile capture, multilingual audio, and what happens to your audio recordings.
Finally, don’t just buy the category. Adopt the behavior. Build your meeting template even if you delete the app. That template will pay for itself before the call ends.
