Your customers have questions.Whoever doesn't answerloses the salethe sale.
askVela answers in seconds with your real products, prices and order data — and takes customers straight to the right product. Around the clock.
7 % of measured conversations ended in an order · 1,341 conversations analysed
Enter your store address — your customers ask the questions in a moment.
Questions turn into orders.
No averages, no extrapolation: four numbers from the production database, each with the sample it came from. The two on the right are the interesting ones — they show what ends up on the invoice.
2026-06-03 – 2026-08-21
2026-06-03 – 2026-08-21
2026-06-03 – 2026-08-21
2026-07-13 – 2026-08-21
How these numbers were produced
All values come from a query against the production database on 21 Aug 2026. Our own workspaces, test instances and the showcase environment are excluded — only third-party companies running the widget on their own site were counted.
- 2.7 seconds to reply
- Gap between a visitor's message and the AI answer that followed it, across 8,588 answer pairs. Gaps above five minutes are excluded: the visitor had left the page, so we'd be measuring their coffee break. Median 2.7 s, 90th percentile 5.4 s.
- 86 % of chats without a human
- 2,072 of 2,408 conversations in which a visitor wrote and the AI answered needed no teammate message at all. Deliberately not called “resolved”: what is measured is that nobody had to take over — not that the customer was happy.
- AI answers sent
- Count of messages the AI has sent in third-party stores — no average, no extrapolation. This is the one tile that doesn't stand still: it checks the current figure every 30 seconds and counts up when something has actually been added. Between two checks there is deliberately no timer ticking along for decoration. Measured, about 160 answers accumulate per day, roughly one every nine minutes — so most of the time you'll see the counter sitting still.
- €9,790 in revenue from customers after chatting
- 437 orders from 289 customers between 13 July and 21 August 2026, €22.40 per order on average. Every order a customer places from their first conversation with askVela onwards is counted — including later ones, including without a new conversation. What this explicitly does NOT say: that askVela caused this revenue. Someone who asked once and buys again three weeks later might have done so anyway; without a control group the two cannot be separated. Counted more narrowly — only orders tied directly to a single conversation — €5,535 across 313 orders remains, €17.68 per order. Both figures are measured; they simply answer different questions.
- And does it pay for itself?
- Across the paying stores in the same period: €341.80 in subscription fees against €2,615.85 in revenue from orders with a directly attributed conversation — roughly €7.60 for every euro spent. The average hides the spread, though: most stores are well above, some below. That is exactly why the guarantee sits below it.
What's missing is missing on purpose: we haven't cleanly measured how far returns drop — the number of processed refunds is still too small to carry a claim. The moment it does, it goes here.
2026-08-21
What this means in your store.
Two inputs are enough. The maths uses the rates measured above, not wishful ones.
A mid-sized store with 20,000 visitors gets roughly 400.
The only assumption in this calculation — set it to your number.
400 conversations × 7 % order rate × €60 order value
86 % of conversations need nobody from your team
At €99 (Growth) per month, askVela pays for itself from 24 conversations a month.
Try it on my storeA model, not a promise. Order rate (94 of 1,341 conversations since 14 July) and deflection rate (2,072 of 2,408) are measured; order value and handling time are yours. The 7 % is deliberately conservative: counting all 235 attributed orders against those same 1,341 conversations would give 17 %, but some of them belong to conversations that began before the measurement window. And whether a customer would have bought anyway without the chat is something this calculation cannot know.
Buying advice, not FAQ search.
The customer describes a problem, not a product. askVela explains the reason, recommends on that basis — and names the page in your store the answer came from.
Answers from your data, with the source attached.
askVela answers from your catalogue, your articles and your order data, and shows what an answer rests on. When that basis isn't enough, it hands over to your team instead of guessing.
What stores use askVela for.
Sell more
The AI spots buying intent, recommends from your real catalogue and shows the product as a clickable card with image, price and availability.
See product recommendationsAnswer more
Shipping, returns, sizes, compatibility, order status: the same questions over and over, handled automatically — at night and on weekends too.
See the knowledge baseLose less
A shopper told the right size and the right model before buying sends fewer parcels back. Refunds run through a reviewed proposal rather than a snap decision.
See refundsFrom store address to an agent that answers.
Connect your store
Shopify via app, WooCommerce via plugin, everything else with a script tag. No developer needed.
askVela reads your store
Products, prices, availability, shipping and return rules land in the workspace as a searchable basis.
Put the widget live
You set the tone, the colours and when to hand over. From then on askVela answers questions and you see every one of them in the inbox.
Why an FAQ bot can't do this.
The difference isn't the language model. It's what the agent has access to.
No video. No mockup. The real dashboard.
One click creates your personal sandbox workspace – with real conversations, real AI and every setting. You click through exactly the product your customers get.
- No sign-up
- 60 minutes, all yours
- Deletes itself
If it doesn't pay for itself, you get your money back.
We don't just claim askVela pays for itself — we stand behind it. If after 60 days your dashboard shows less attributed revenue than you paid us in that time, we refund the full fee. No questions, no justification, no retention call.
Start with my store- Measured with the purchase attribution in your own dashboard — the same figure you can check yourself at any time.
- That requires the widget to be live and the store connected (plugin or app). Without a connection nothing can be attributed.
- The window is the first 60 days after the free trial. What is refunded is the subscription fee paid in that period.
This guarantee isn't free for us — there are stores where it applies. We offer it anyway, because we would rather refund than keep charging someone it isn't working for.
Every feature in every plan.
No per-seat pricing, no fee per resolved case. The only thing that differs between plans is the number of AI replies per month — and it's stated openly in the table rather than buried in the footnotes.
- Unlimited agents
- 1,000 AI replies/month
- AI bot + knowledge base
- Shop integration + order lookup
- Product recommendations
- EU-hosted, GDPR/DPA
- Everything in Starter
- 4,000 AI replies/month
- AI actions: refunds & cancellations
- All messaging channels
- White-label, webhooks, reviews
- Priority support
- Everything in Growth
- 12,000 AI replies/month
- Premium AI routing
- Multiple shops + broadcasts
- AI reply packs available
- Onboarding + priority support
- Quota and price by agreement
- SSO & custom configuration
- SLA & a named contact
- Personal onboarding
If the quota ever falls short, a pack of 1,000 extra replies costs €25. Your plan doesn't change on its own.
Every feature at a glance
Each feature has its own detail page — explore them and see exactly what it does.
See it on your own store.
Enter the address and ask askVela something your customers actually ask. If the answer is no good, you've lost a minute — and we have a problem worth knowing about.
No signup · no credit card · the demo deletes itself
Everything else that's in there.
In case you're looking for one specific capability. For the buying decision, the part above is enough.