The FAQ Page Is a P&L Item, Not a Design Detail
Cross-border sellers spend real money dragging strangers to a product page and then let the last ten feet of the funnel rot. I’ve watched operators burn four figures a day on Meta traffic to a Shopify PDP whose FAQ block was written once during launch week, from guesses, and never touched again. Meanwhile the actual questions — “does this fit a 220V outlet,” “will customs seize this,” “how long to Germany really” — go unanswered, and the visitor leaves. That’s not a content problem. That’s a conversion and returns problem wearing a content costume. So when a tool shows up that treats FAQ as a living, measurable asset rather than a static theme block, I pay attention. That’s the lens I’m using on Actually Asked, launched by maker Petar.
What Problem It Actually Solves (And Why Static FAQs Are Bleeding You)
The pitch is deceptively simple: a FAQ section that lives inside your page, in your own fonts and colors, where visitors can open an entry, ask their own question beneath it, or highlight any sentence on the page and ask about that. Answers are generated only from your own pages, products, policies, and files — not from the open web. When the knowledge base can’t cover something, the visitor can leave an email so you capture the demand instead of losing it silently.
Read that last clause again. “If those don’t cover it, the visitor can leave their email for you.” For a Shopify operator, that’s a lead-capture mechanism disguised as customer service. For an Amazon seller, it’s a preview of the questions your listing copy is failing to preempt.
The genuinely interesting mechanic is the self-improving loop. Per the maker’s own reply in the thread, when a few different people ask the same thing within a month — however they word it — the system groups them, writes one clean question with an answer drawn from the learning materials, and publishes it to the FAQ automatically. Ask it on a product page and it surfaces on that product. Anything you’ve edited, pinned, or removed stays exactly as you left it. That last guarantee is the one that separates this from the graveyard of “AI writes your content” tools that operators abandon after the second hallucinated policy answer.
The Monday email is the actual product
A commenter flagged it and I agree: the weekly email listing what your pages couldn’t answer is the sharpest idea in the whole launch. Most sites never see the questions people had before they left. Every unanswered question in that digest is a content gap, a pre-purchase objection, or a future return. That’s a defect log for your product page, delivered on a schedule.
How It Differs From What You’re Already Running
Let me place it against the incumbents honestly, because “AI FAQ” is a crowded shelf.
Against Zendesk / Gorgias / Intercom: Those are support desks. Their entire architecture assumes a human or bot conversation in a chat bubble layered over your page. The maker’s own framing is the sharpest line in the launch: “Chat bubbles cover the page and feel like a support desk.” For a DTC brand, a chat bubble signals “something is wrong, talk to a human.” A native FAQ block signals “this brand anticipated my question.” Different psychology, different conversion behavior. Gorgias and Zendesk are where you resolve problems; this is aimed at preventing the contact in the first place.
Against Tidio / Chatwoot: Same category critique — chat-first, not content-first. They capture conversations; Actually Asked captures questions as structured content and feeds them back into the page.
Against your theme’s built-in FAQ block or a Shopify accordion app: Those are static. They don’t learn. The gap between a static accordion and a self-updating FAQ is the gap between a printed manual and a support engineer who reads every ticket.
Against generic AI chatbots trained on the open web: This is the differentiator I’d press hardest. Answers come only from your pages, products, policies, and files. For cross-border sellers, that’s not a feature — it’s a liability shield. A chatbot that invents a shipping time or a warranty term on a German storefront is a legal and CSAT problem. Grounding in your own corpus is the correct architecture.
Why Amazon sellers should care more than Shopify ones
Shopify operators get the obvious win: a theme block on the Shopify App Store, native styling, no dev time. But Amazon sellers — who can’t install anything on Amazon Seller Central — should be reading this as a research instrument, not a deployment. The Monday digest tells you what your PDP copy, A+ content, and bullet points failed to answer. Every grouped question from the FAQ is a candidate bullet rewrite or a Helium 10 keyword you’re missing. You can’t embed the widget, but you can absolutely run the same question-mining loop manually against your own listing’s Q&A section and review corpus. The tool is a template for a process.
What Cross-Border Sellers Can Borrow From This
Even if you never install it, three transferable ideas:
Treat unanswered questions as a KPI. Most operators track conversion rate and return rate but never track “questions we failed to answer.” The Monday digest concept — a recurring report of gaps — is worth building yourself even with a spreadsheet and a manual review-mining pass.
Ground every AI answer in your own corpus. If you’re running any AI customer service layer, the “answers come only from your pages, products, policies and files” constraint is the rule you should enforce. Open-web-trained bots are a returns and compliance grenade for cross-border.
Make human edits sacred. The “anything you edit stays the way you wrote it” guarantee is the trust mechanism that makes automation tolerable. Borrow the principle: automate the draft, never the final word on policy, pricing, or shipping promises.
Where the math breaks
The free plan gives 50 AI answers a month on one site, no card needed. For a store doing real volume, 50 answers is a rounding error — likely a single busy afternoon. That’s a deliberate funnel: the free tier is a demo, not a plan. Before you commit, model your actual monthly question volume against the paid tier’s pricing, which is not disclosed in the launch material. If you’re a high-traffic DTC brand, the cost-per-answer math is the whole decision, and you can’t do it from this page.
Where My Judgment Says It Falls Short
Highlight-to-ask is half-built. A commenter raised the exact right concern: highlighting UI text in a button or navigation and getting a popup is odd. The maker’s answer is candid — selections only count between 3 and 300 characters, are ignored inside form fields and editable text, and never auto-open the popup (a small “Ask about this” button appears instead). But he concedes it: “What it doesn’t do yet is skip buttons and navigation, or let you mark parts of the page as off-limits.” For a clean storefront that’s a minor annoyance. For a complex PDP with variant selectors and configurators, it’s a UX papercut you’ll field complaints about.
The multi-platform story is impressive but shallow where it counts. It ships as a Shopify theme block, a Framer plugin, a connector in Claude, and works on WordPress, Webflow, or any site with two lines of HTML. It’s also manageable from Claude, ChatGPT, Cursor, or Codex over MCP. That’s a lot of surface area. Breadth like this usually means the deepest platform — Shopify — gets the most polish and everything else is a thinner integration. If you’re on Webflow or a headless stack, validate the styling fidelity before you trust the “your own fonts and colors” promise.
The auto-publish loop needs a leash. The system groups repeated questions and publishes answers automatically. For low-stakes questions, great. For anything touching shipping timelines, duties, or warranty, auto-publishing a synthesized answer is a risk. The manual-override guarantee protects you if you catch it — but the default is autonomy. I’d want a “review before publish” toggle for policy-adjacent categories, and the launch material doesn’t confirm one exists.
No mention of multilingual behavior. This is the glaring omission for a cross-border audience. A German shopper asking in German, a Japanese shopper asking in Japanese — does the FAQ answer in their language, and does the grouping logic collapse “does this ship to X” across languages? The launch says nothing about localization. For a single-market US store, irrelevant. For anyone running EU or APAC storefronts, it’s the first question I’d ask.
What I’d Watch / Test Next
Three concrete moves this week, no matter your platform:
Run a manual version of the Monday digest. Pull your last 90 days of Amazon buyer messages, Shopify contact form submissions, and pre-purchase chat transcripts. Group the repeated questions. That list is your FAQ rewrite backlog and your next A/B test on PDP copy — you don’t need the tool to start.
If you’re on Shopify, install the free tier and instrument it. Fifty answers a month is enough to see whether visitors actually engage the inline block versus your existing chat widget. Watch the email-capture rate on unanswered questions — that’s the metric that tells you whether this is a support tool or a lead-gen tool for your store.
Get pricing clarity before you scale. Email the maker and ask for paid-tier pricing and per-answer costs at your volume. The free plan’s ceiling is documented; everything above it is not disclosed. Don’t build a workflow on a tool whose unit economics you can’t model.
The strategic takeaway: the FAQ is a living asset, and whoever treats it that way — tool or not — wins the last ten feet of the funnel. That’s the idea worth stealing, whether or not you install the app.
