Frequently asked questions

Pickup recovers abandoned checkouts with a real phone call. When a shopper leaves without buying, your AI voice agent calls them within minutes, handles the objection in a natural conversation, and emails their checkout link (and a discount, if you offer one) so they can finish in a couple of taps. Every call is logged, transcribed, and attributed to recovered orders.

Three steps, all in the app: (1) set your calling hours in Settings, (2) review your recovery discount and email sender, (3) place a test call to hear your agent. Cart-recovery is on by default once a cart is abandoned and you have credits. There's no switch to flip.

Use the "Call me now" card on your Dashboard (it's in the getting-started checklist): enter your own phone number and your agent rings you within about a minute, so you hear exactly what a shopper hears. A test call needs no abandoned cart, uses none of your credits, and never creates a discount code. You get 2 test calls per store per 24 hours and each one is capped at 5 minutes; if you use both, you're simply asked to try again tomorrow and your live recovery calls are completely unaffected. A number that has opted out of calls can't be used for a test, even your own.

Yes, in Settings → Email sender. Until you connect a domain, recovery emails go out from the shared Pickup address with your shop's name as the sender, so they still read as coming from you. To send from your own domain: add it, add the DNS records we list for it, then click Verify. DNS can take a few hours to propagate. If a check comes back empty, that's what happened; wait and verify again. Once verified you can switch back to the shared address at any time.

Because something went right. Pickup only asks at genuine success moments: your first attributed recovered order, roughly every fifth recovered order after that, when your all-time return on Pickup passes 2×, and after a knowledge-base import succeeds. The dialog itself is Shopify's, and Shopify caps how often it can be shown, so you won't be pestered. Declining costs you nothing and changes nothing about your account.

Credits are your calling wallet: 1 credit = US$0.20. Calls bill by the minute, rounded up to the next whole minute, exactly the way phone carriers bill (a 64-second call counts as 2 minutes). The per-minute rate depends on the destination country. It's US$0.20/min for the US, Canada and the UK, and more for destinations where mobile networks charge more to reach. The Call rates page lists every country. You're only charged for connected talk time; unanswered calls cost nothing. There's no SMS. Pickup is voice-only.

A typical recovery call runs a few minutes, so figure roughly 3–5 credits per connected call to the US, Canada or the UK. Where the local mobile network charges more to reach, a call costs proportionally more: a German or Gulf call can run 2–3× that, and the Call rates page lists the per-minute price for every country. The app shows an "≈ N recovery calls" estimate next to your balance, based on US rates, so treat it as a guide rather than a promise if you sell mainly outside those markets.

Your monthly plan includes a credit allotment that's granted at the start of each billing period. Paid plans renew monthly. Top-up credits you buy separately never expire and stack on top of your plan credits.

On the Billing page you can buy a one-time pack of credits any time, on any plan including Free: either one of the fixed packs or a custom amount, anywhere between the minimum and maximum shown on the page. You approve the charge through Shopify, and the credits land in your balance and never expire. Top-ups are separate from your plan's monthly allotment, so buying credits doesn't change your plan, your weekly cap or your features.

You can change plans any time on the Billing page. An UPGRADE applies immediately: Shopify prorates the change (you pay the difference for the remainder of the cycle, not a fresh full charge) and your credits are topped up to your new plan's monthly allowance. A DOWNGRADE applies at your next renewal: you keep your current plan, its features and its credits until the period you've already paid for ends, then move to the lower plan. CANCELING works the same way: you keep your plan until the period ends, then move to Free. In every case, the credits already in your balance stay with you.

You do end up with your new plan's full monthly allowance. We simply add the difference rather than a second full grant. Each plan includes a set number of credits per billing month, and you had already received your previous plan's credits for the current month. When you upgrade, we top your balance up by the gap between the two plans so you reach your new tier's full monthly amount. For example, moving from a plan that includes 30 credits to one that includes 150 partway through the month adds 120 credits, bringing you to the 150 your new plan includes, not 180. This keeps things fair on both sides: you receive the complete allowance for the tier you're now paying for, without a single month's credits being granted twice. At your next renewal your balance refreshes to the new plan's full allowance, and any top-up credits you've purchased always remain on top.

A downgrade takes effect at the end of your current billing period rather than the moment you request it, so nothing is taken away early: you keep your current plan, all of its features, and its credits right through to your next renewal date. Because you continue to receive the full plan you have already paid for until that date, there is no partial refund for the remaining time; you are getting exactly what you paid for. At renewal you move to the lower plan and are billed its rate from then on, and any credits already in your balance stay with you. The Billing page shows a clear "switches to [plan] on [date]" note so you always know precisely when the change takes effect.

The Billing page shows a "Downgrade scheduled" notice naming the date and the plan you'll move to. Until that date literally nothing changes: you keep your current plan, its features and its credits. If you opened Shopify's approval screen but closed it without approving, no downgrade was ever booked. The notice is just stale, and you can dismiss it with the Dismiss button. If you did approve it and have changed your mind, move back up whenever you like: an upgrade applies immediately. A notice that lingers past its own date without the plan ever changing clears itself.

Each plan includes a weekly cap on how much recovered revenue Pickup will chase before pausing new calls, a guardrail so spend tracks results. It resets every Monday (00:00 UTC). If you hit it, new calls pause until the reset or until you upgrade. It's separate from credits: you can have credits but be capped, or vice-versa. You'll know you've hit it in three places: a warning banner on your Dashboard, a matching banner on the Billing page naming the cap in your store's own currency, and a "Weekly recovered-revenue cap reached" label on the affected rows in your Calls log.

Every plan sets four things: the credits granted each billing month, the weekly recovered-revenue cap, how many times your agent may call one shopper who didn't answer, and how many calls may run at the same time. Here is each plan, from the bottom of the ladder up. Free: 5 trial credits, US$40 recovered per week, 1 call per shopper (no retry), 1 call at once; Lite: 10 credits a month, US$80 recovered per week, 1 call per shopper (no retry), 1 call at once; Starter: 30 credits a month, US$300 recovered per week, up to 2 calls per shopper, 2 calls at once; Pro: 70 credits a month, US$600 recovered per week, up to 3 calls per shopper, 3 calls at once; Growth: 150 credits a month, US$1,000 recovered per week, up to 3 calls per shopper, 5 calls at once; Business: 320 credits a month, US$1,500 recovered per week, up to 3 calls per shopper, 10 calls at once; Premium: 700 credits a month, US$3,500 recovered per week, up to 3 calls per shopper, 20 calls at once; Scale: custom credits, cap and simultaneous calls, built around your volume. Credits and the weekly cap are independent: you can have plenty of credits and still be capped, or the other way around. You can move between plans any time on the Billing page, and your Billing page always shows the limits for the plan you're actually on.

Each plan sets how many recovery calls may be live at once: one at a time on the entry plans, and more on every step up. Going over the limit never loses a cart: the extra calls wait in the queue and dial the moment a line frees up, in the order they arrived. It only matters when carts arrive in bursts (a flash sale, a big campaign, a newsletter send), where a higher plan clears the queue faster and reaches shoppers while the cart is still fresh.

No. Every call is hard-capped at whatever your balance can pay for. That means a thin balance doesn't block the call. It cuts the conversation short, sometimes right as your agent is closing the sale. It's the most expensive kind of quiet failure, because a truncated call is logged as a perfectly ordinary completed one and nothing looks broken; you'd just watch your recovery rate slide. That's exactly what the credit-runway warnings on your Dashboard exist to prevent, so treat one as a reason to top up now rather than later.

Because runway is priced by destination, not by a flat number of credits: the same balance can be many minutes to the US and less than one whole minute to an expensive mobile network. You'll see one of two banners, never both. The urgent one appears when carts are already queued that your balance can't fully fund: it names the priciest country in that queue, how many whole minutes your balance buys there, and how many queued calls would be cut short (or, below one minute, not placed at all). The heads-up one appears when nothing is at risk yet but your balance is down to fewer than three full calls at the priciest destination you've dialed recently. Both link straight to the rate for that country so the price makes sense. Neither shows at a zero balance; the out-of-credits banner covers that.

Everything goes through Shopify's billing: your subscription and any top-ups appear on your regular Shopify invoice using whatever payment method you have on file. Pickup never handles your card directly.

The Call rates page lists every destination, with a region picker so you can find your markets quickly. The price shown is all-in: the carrier's rate to reach that country plus the AI that places and speaks on the call. Nothing else is added on top. Inbound calls and abandoned-checkout alerts are always free, and you only ever pay for connected talk time. Rates track what carriers charge us and can move when they change theirs. A destination that isn't listed bills at the fallback rate stated on that page. If you're planning a campaign into one, talk to us first and we'll quote it properly. All prices are in US dollars.

Pickup has a ladder of monthly plans you can switch between any time on the Billing page: Free (free trial credits); Lite (US$4.99/mo, 10 credits); Starter (US$29.99/mo, 30 credits); Pro (US$49.99/mo, 70 credits); Growth (US$99.99/mo, 150 credits); Business (US$199.99/mo, 320 credits); Premium (US$499.99/mo, 700 credits); and Scale, a custom contact-us tier for high volume. Each step up adds more credits, a higher weekly recovered-revenue cap and more calls running at once, and the higher tiers also add no-answer retries, call recordings, AI deep-analysis, sending recovery emails from your own domain, spoken voicemails and CSV export.

Scale is Pickup's top, white-glove tier for stores recovering at high volume. Rather than fixed limits it's built around your volume: custom credits and caps, a dedicated account manager, 24/7 priority support, custom call scripts and recovery strategy, and a direct Slack/WhatsApp channel. It isn't a self-serve card you click to buy: tell us your volume and we tailor it to you. To start, reach out through the Scale form (https://pickup.exil.ai/scale) or email pickup@exil.ai and we'll set it up.

The Referrals tab in the left-hand menu. Your store gets its own share code the moment you install (it has your store's name in it, so it's recognizable when you send it), and it stays the same for as long as Pickup is installed. (If a store uninstalls and Shopify erases its data, a fresh install can mint a new code; anyone who already redeemed the old one keeps their referral.) Copy it with the Copy button and pass it to another store owner: they install Pickup from the Shopify App Store and type your code on their own Referrals page. The code can't be baked into an install link, so it does have to be typed there.

Both sides earn, and each side is paid on its OWN plan: Free earns 5 credits, Lite earns 10, and Starter and above earn 20. There is deliberately no single shared number. If you're on the free plan and refer a store on a paid one, you earn the free rate and they earn theirs, and neither of you can change the other's amount. Your own rate is shown on the Referrals page, along with what the top rate would be if you moved up.

Nothing is paid when a code is applied. Credits land on both sides only once the store you referred has completed, and been charged for, its first real recovery call to one of their own shoppers. A free demo call doesn't count, and neither does a call that never connects; that one is refunded in full, so it costs nothing and converts nothing. Until it happens, your Referrals page shows them as "waiting on a first call". When it does, the store you referred is paid immediately. Your own side is paid on the next daily payout run (normally within a day) unless you've already had your 100 payouts for the month, in which case it joins the queue and is paid from the 1st, oldest first. Nothing is lost either way, and we email you when credits land.

You can apply a friend's code at any point in your life on Pickup: there's no eligibility window and no deadline, whether you installed today or last year. But a store can only ever apply ONE code, and it can't be changed or swapped afterwards, so be sure it's the right one before you hit Apply. Your own share code is a separate thing entirely and is always active, whether or not you've ever used someone else's.

The page tells you which of these it was: the code doesn't match any store (it's easy to mistype, so check it with your friend); it's your own code, or it would loop back between stores that have already referred each other; you've already used a code (one per store, ever); development stores can't apply a code, so use a live store; your store isn't live yet, in which case remove the password page and try again; too many incorrect codes in a row, which pauses attempts for a while; or we couldn't verify your store with Shopify just then (that one is temporary, so try again in a minute). How much the code's owner has already earned is never a reason: a code keeps working for every store, and you're always paid your own bonus in full.

No, and that's deliberate. A store that used your code is another merchant's business, so we never share their name, their domain, or when they joined. "Your referrals" is a totals-only view: referrals paid out, credits earned, a chart across the date range you pick, your all-time totals, how many are waiting to be paid, and how many are still waiting on a first call. The only date tied to an individual referral is your own: the day you applied someone else's code.

For referrals from then on, yes. Each side's amount is locked in the moment the code is APPLIED: whatever plan you're on right then is the rate that referral pays, for good. So moving up BEFORE you share your code (or before you apply a friend's) is what lifts your rate; upgrading afterwards to cash in referrals that are already sitting and waiting won't re-price them. It works the same way in reverse: dropping to a lower plan later never reduces a referral you'd already banked at a higher rate.

There are two limits, and both only ever apply to what YOU earn: never to whether your code works, and never to what the store using it receives. First, up to 100 referrals are paid to you per calendar month; anything above that isn't lost, it keeps its place in the queue and is paid in a later month, oldest first. So a code that's redeemed 500 times pays you 100 a month for about five months, while all 500 of those stores are paid their own bonus straight away. Second, a store can be paid for at most 1,000 referrals in its lifetime; past that you stop earning, but your code still works for everyone else and every store using it is still paid in full. Your Referrals page shows how many payouts you've had this month, how many are waiting, and when the next batch lands.

Bonus credits on top of what each referral already pays you, at seven named rungs: Advocate at 10 referrals (25 credits), Champion at 25 (50), Ambassador at 50 (100), Elite at 100 (200), Legend at 250 (450), Icon at 500 (800), and Hall of Fame at 1,000 referrals (1,500 credits). The Referrals page shows every rung, which you've unlocked, and exactly how many more referrals stand between you and the next one. Only referrals that have actually paid you count towards them; one that resolved at zero because you're past your lifetime cap moves you no closer. Each rung pays once ever per store, and the bonuses are flat rather than plan-based, so there's nothing to time around an upgrade: the credits per referral stay about the same all the way up the ladder.

Only inside the calling window you set in Settings → Calling rules (e.g. 9am–9pm), in each shopper's own local time where we know it; otherwise your store's time zone. This keeps calls compliant (TCPA) and friendly. Five controls live on that card: the start hour and end hour of the window, how long to wait after a checkout is abandoned before the first call, how many hours to wait before a second call, and what hour a third call lands the next day. The last two only appear if your plan includes retries.

Depending on your plan, the agent retries a missed call up to a few times (one attempt only on the entry plans, more further up the ladder) with a delay between tries; a 3rd attempt lands the next day at an hour you choose. It never exceeds a hard cap of 3 attempts per shopper on any plan, including the top ones. No spam. Someone who answers and declines is never called again. Your Settings page shows how many attempts your plan gives you.

No. Pickup places one recovery call per cart, for that cart's entire lifetime. The rule is keyed to the cart rather than the shopper, so even if the same checkout is reported to Pickup several times while the shopper edits it, it can still only ever produce one call. Automatic no-answer retries are a different thing: they re-use that same call rather than starting a new one, and they're bounded by your plan's attempt limit. If the same shopper abandons a genuinely new cart later, that's a new cart and can be called about on its own.

It means that shopper had already been dialed as many times as your plan allows within the last 24 hours, so Pickup didn't call them again. Three things about it surprise people. It counts across ALL of that shopper's carts, not one: someone who abandons twice in an afternoon still only gets your plan's allowance of calls in total, because being rung repeatedly by the same store in one day is what makes a store look like a spammer. It counts no-answer redials too, so a cart that rang out twice has used two of them. And it's a rolling 24 hours from the last call rather than a calendar day, so it clears gradually rather than all at once at midnight. There's no separate setting for it: the number is your plan's per-shopper attempt allowance, shown on your Settings page, and moving up the ladder raises both together.

If a shopper asks the agent to stop during the call, or unsubscribes from a recovery email, they're added to your Do-Not-Contact list immediately and never contacted again; those automatic entries can't be removed. You can also add numbers yourself in Settings → Do Not Contact: enter the phone number and, if it's useful for your own records, an optional reason ("complained", "legal request").

On Growth and above you can have your agent leave a short, friendly voicemail: Settings → Voicemail, the "Leave a voicemail message" switch. Below Growth that card is locked with an upgrade link and the agent always hangs up quietly instead. Either way the recovery email with the checkout link still goes out, so the shopper can finish from their inbox. A voicemail is its own outcome, not a conversation: it gets its own bar on the Dashboard and its own filter in the Calls log, and it's never counted as "Spoke with customer" (a machine answered, not the shopper), even though the call is still analyzed like any other. On the call itself the next step reads "Reached voicemail", and it only claims the checkout link was emailed if one genuinely went out. Reaching voicemail counts as making contact, so Pickup won't keep re-dialing that cart.

Every call lands in one bucket on the Dashboard breakdown: Connected (the agent reached the shopper), Voicemail (a machine answered; kept separate so voicemails never read as conversations), No answer (it rang, no pickup), In progress (queued or still on the line), Blocked, or Failed. The distinction people ask about most: "Blocked" means we never dialed: the call was screened out before placing it (the shopper left no phone number, there's no consent on file, the number looked like a burner/VoIP line, it's on your Do-Not-Contact list, you'd hit a daily cap or your weekly cap, you were out of credits, or it was outside your calling hours). "Failed" means we DID dial but the call never connected (the number was unreachable, the carrier rejected it, or it timed out). Because of that, the same kind of problem number can show under EITHER bucket depending on whether we got far enough to actually dial it. Two more labels show up on individual rows rather than as buckets: "Recovered", once an order is attributed to that call, and "Closed", for a call you've closed by hand. Closed is treated as a deliberate non-outcome and is left out of every breakdown, so the bars won't always add up to your full call list. To see precisely which calls are in a bucket, open the Calls tab and filter by status.

Each row carries the precise reason for that one call, and they roll up into the Dashboard's buckets: Likely a burner / VoIP number · Unreachable number · No consent on file · Do not call · Outside calling hours · Weekly recovered-revenue cap reached · Daily limit reached · Out of credits · Already purchased · Couldn't reach this number · Timed out · No answer · Hung up early · Couldn't place the call. Two get confused. "Couldn't reach this number" means we dialed and the line never came up, typically a landline, a VoIP line, or an invalid or made-up test number; real mobile numbers connect normally, so a couple of these among test numbers is not a fault. "Timed out" means the call was left hanging with no clear ending and we closed it out.

Two filters at the top of the Calls tab. By status: All, Completed, Voicemail, Failed, No answer, In progress, and "Asked for a human". By date: All time, Today, Last week, Last month, Last 3 months, Last 6 months, Last year. The status options line up exactly with the Dashboard's outcome bars, so a bucket on the Dashboard and its filter here always show the same calls.

As a banner at the top of your Dashboard counting how many customers asked. Clicking Review marks them as seen and drops you into the Calls log filtered to exactly those calls; if a new request comes in afterwards, the banner returns with a fresh count. On the call itself the next step reads "Asked for a human — call them back" with their email or phone right there. It's picked up from what the shopper actually said on the call, and a phrase used in the negative ("please don't call me back") deliberately does not flag, because that's an opt-out, not a request. This is the one follow-up Pickup ever asks you to do by hand, so it's worth doing fast.

A plain-English read of what happened and what, if anything, you should do. Every call gets one. You'll see things like "Recovered — no action", "Reached voicemail", "Asked for a human — call them back", "No answer — retrying tomorrow", "Retried — no pickup", "Couldn't reach this number", "Not interested", "Customer hung up", "Check your checkout" (when the shopper hit a payment or page error worth testing yourself), or "Checkout link + discount code sent" with your actual percentage. It never claims an email was sent when none was: if there was no address on file, or the send didn't go through, it says so and suggests following up by phone instead.

On Growth and above, yes: the Export CSV button on the Calls page downloads your calls and recoveries as a spreadsheet. Below Growth the button is visible but locked and links to Billing, so you can see what you'd be getting before you pay for it.

Deep analysis is a second, closer read of a single conversation. Open any call and use "Run deep analysis" (or "Re-run") in the detail panel. It's a Pro-and-above feature, as are call recordings: below Pro the button is a lock that links to Billing, and the recording area says plainly that recordings are a Pro feature. A call needs a transcript before it can be analyzed, so a call that never connected has nothing to work with. If you're already on Pro and it reads "Temporarily unavailable", that isn't an upsell; the analyzer is briefly down and the button comes back on its own.

You control call volume with a few levers rather than a single on/off switch. There is no master toggle, and no switch on the Agents tab: your agent is always on by design. Narrow your calling hours in Settings and calls only go out inside that window (in the shopper's own local time). Your credit balance is the spend control: when you're out of credits, no new calls go out, so how much you top up sets the pace. Each plan also has a weekly recovered-revenue cap: hit it and new calls pause automatically until it resets Monday or you upgrade. To stop calling specific people, add them to Settings → Do Not Contact (a shopper who opts out on a call is added there automatically). Any call already in progress always finishes naturally.

If you enable it in Settings, the agent creates a unique, single-use discount code on the spot near the end of the call and emails it to the shopper, a closing nudge. It's a one-time code that expires within the hour, so it's a genuine reason to finish now. You choose the percentage (up to 90%).

Yes. Set "Never offer a discount on calls" in Settings and the agent won't create a code or mention any of your saved codes. Recovery then relies on the call + the checkout link alone.

Yes, save up to 8 of your own standing codes in the Knowledge base, each with a short description of when it applies. Every code is checked against Shopify when you save it, so the agent can never offer one that's invalid or expired. That single list does two jobs at once: it's what the agent is told exists, and it's the allowlist the follow-up email is checked against, which is why the agent can never email a code it wasn't given. The one-time code it creates during a call is separate from these.

Lilia, your cart-recovery agent, is fluent in English, Spanish and Arabic. The language a call OPENS in is chosen from the script the shopper typed their own details in at checkout: Arabic characters in their name or their address and the call opens in Arabic, in a native Arabic-speaking voice rather than her English one; anything else opens in English. That matters because an Arabic-first shopper who hears a foreign-language opening tends to hang up before anyone can switch. From the first reply onwards the agent simply speaks whatever the shopper speaks, and the follow-up email goes out in the same language. There's nothing to configure.

It answers from your store's own knowledge base: your products, prices, and policies (shipping, returns), plus any custom Q&A you add. On the call it looks things up live so it answers from real data instead of guessing, and offers to email the shopper if it doesn't know something.

Yes. If a shopper says they've already paid, or asks where their order is, the agent looks them up in your store's real order records during the call (by email first, then by phone number) and reads back what it finds: order number, date, payment status, fulfillment status and total, for up to their three most recent orders from the last 30 days. Anything older than that isn't treated as "this purchase", so a repeat customer's order from months ago is never confirmed as payment for the cart you're calling about. If there's no matching order, or it couldn't check just then, it says so plainly rather than guessing.

One today: Lilia, your cart-recovery agent. She's the one who phones your abandoned checkouts, and she's shown by name on the Agents tab along with the languages she speaks; every call in your Calls log is attributed to her too. She's always on: there's nothing to switch on, and no way to switch recovery off from that page (calling hours, your credit balance, the weekly cap and your Do-Not-Contact list are how you control call volume). The same tab shows the specialists we're building next as a roadmap: Anna, who wins back lapsed customers; Ellie, who collects post-purchase feedback; Maya, who answers and qualifies inbound calls; Emma, who recovers failed payments; and Hannah, who books and confirms appointments. None of them is live yet and none can be switched on today; they'll appear alongside Lilia as they're ready.

Two ways, and most stores use both. Import from your store pulls in your active product catalog and your published refund and shipping policies in one click. That's available on EVERY plan, including Free; it is not gated behind a tier. Then, in the Knowledge base tab, hand-add the things an import can't know: sizing advice, delivery times, your returns window, and any custom Q&A. You can edit imported products in place, and your edits survive future imports, while products you haven't touched keep tracking your store. The more you add, the better the agent answers, and the better it sells.

Products and policies. When it finishes you get a count of exactly what landed ("Imported 12 products and 3 policies") rather than a vague success message. It can take up to a minute, so leave the page open while it runs. Re-import any time your catalog changes; it's safe to run as often as you like, and your own edits to imported products are preserved each time. A result of 0 items isn't an error: it means your store currently has no active products and no published refund or shipping policies to read, so publish those in Shopify and import again.

Yes, with one sensible limit. Pickup recognizes a storefront password gate and skips it, so the "this store is password protected" page is never learned as though it were your product copy, which is exactly what would otherwise end up being read out on a call. Your products and your published policies still import from your store's own records. Once you launch, run Import again to pick up everything that lives on your public pages.

In the Knowledge base tab choose "Add a new product," give it a name, then in Details write each fact as a short sentence separated by periods: price, sizes or variants, colors, materials, shipping, anything a shopper might ask. Order doesn't matter and there's no fixed template; just keep facts short and period-separated, e.g. "$128. Sizes XS–XXL. Grey, navy, forest. 100% merino wool. Ships in 3–5 days." The preview shows exactly what the agent will say. The more you add, the better it answers and sells.

Five, all to your store owner's address, and that's the whole list. There's no drip campaign. A heads-up when your credit balance is running low. A note when you hit your weekly recovered-revenue cap. A short celebration when you cross a recovery milestone. A weekly "Your week with Pickup" digest. And a confirmation when referral credits land in your balance. These are separate from the recovery emails your agent sends your shoppers, which go out on every call.

The milestone and weekly-digest emails carry an Unsubscribe link at the bottom; confirming it stops product updates and digests permanently. Essential account notices (a low balance, a cap hit, referral credits landing) keep sending, because they're about your money and your account rather than marketing. There is no in-app toggle for this; the link in the email is the only route, which also means we can't be talked into unsubscribing someone else's store. Opening the link doesn't unsubscribe you on its own: you have to click the button on the page, so a mail-security scanner following links can never unsubscribe you by accident.

No. You get one per low patch, not one per day. It sends when your balance covers fewer than three full calls at the priciest destination you've been dialing recently, then goes quiet. It only re-arms once your balance is healthy again, so topping up and later running low earns you a fresh warning, while sitting low doesn't earn you a stream of reminders.

A short celebration when your lifetime call-driven recoveries cross 5, 10, 20, 25, 30, 50 and 100, and then every further 100 after that. Each one fires once, even if you jump several at a time. It counts recoveries your agent actually drove, not organic orders. It's marketing mail, so the Unsubscribe link in it stops it for good.

Once a week, and only if something actually happened in the previous 7 days; we never send an empty digest. It carries the carts recovered and what they were worth, in your store's own currency; if nothing has converted yet but calls went out, it reports the calls instead so you can see the work happening. It's marketing mail, so the Unsubscribe link stops it.

Because new calls had just paused. It sends at most once a week, when your weekly recovered-revenue cap is reached: it names what Pickup recovered for you that week and the cap on your current plan, then points at the next tier up with its price, its cap and its credits, since raising the cap is the only way to keep calling before the week resets. On the top self-serve plan there's no next tier to offer, so it invites you to talk to us about a custom Scale plan instead. It's an account notice rather than marketing, so it isn't covered by the Unsubscribe link.

Settings → "Feedback & feature requests". Pick what kind of note it is (feature request, bug report, suggestion or other), write up to 4,000 characters, and it goes straight to the team's inbox. There's a per-store daily limit purely to protect the send path; if you reach it you'll see a friendly note and can send the rest tomorrow. For anything urgent, email pickup@exil.ai directly.

Yes. Data is encrypted in transit and at rest, each store's data is isolated, and Pickup only uses the minimum needed (name, email, phone) to place the recovery call and email. Call data is not kept indefinitely: transcripts, summaries, analyses and the call audio are erased after a fixed retention window (30 days by default), and at the same time the abandoned-cart record's customer name and recovery link are scrubbed too. We honor Shopify's data-deletion requests. See the Privacy Policy for full detail.

Usually one of: (1) you're out of credits (top up on Billing); (2) it's outside your calling hours; (3) the abandoned shoppers didn't leave a phone number at checkout, so there's no one to call; (4) you've hit your weekly recovered-revenue cap (calls resume Monday or on upgrade); (5) those numbers are on your Do-Not-Contact list, or they didn't validate (a burner, VoIP or mistyped number); or (6) those carts were already called (Pickup places one recovery call per cart). What it is never is an off switch: your agent is always on and there's nothing to enable, so a missing toggle is not the explanation. The Dashboard's Call-outcomes card names the biggest blocked reason inline, and the Calls tab shows the exact reason on each call.

All-time revenue recovered by calls, set against all-time spend on those calls: the credits actually consumed, net of anything refunded. Both sides are lifetime figures, which is why the multiple doesn't move when you change the date range above it. Revenue from shoppers who came back on their own never counts towards it. We don't hide an unflattering number either: as long as the ratio is at least 0.1× you'll see it, even when it's below 1×. The badge only disappears when there's too little spend for a multiple to mean anything at all.

"Blocked" means the call was stopped before we dialed, and the bar names the biggest single reason inline: out of credits, outside call hours, do-not-call list, invalid numbers, daily limit reached, or stopped before dialing. When blocking is both the largest bucket and a real share of your activity, you also get a banner spelling it out and, for credits, a link to top up. A handful of blocked carts is normal, since some shoppers leave a bad number or none at all; a large share is worth acting on, and the label tells you which lever to pull.

Today, Last week, Last month, Last 3 months, Last 6 months and Last year. It applies to the KPI cards, all three charts and the call-outcomes breakdown. Two things deliberately ignore it: the "Recovered with Pickup so far" figure at the top and anything labeled all-time, which are always lifetime numbers so they can't be made to look better by picking a good week. The Referrals page offers the same six ranges.

By the strongest signal available, in this order: the unique discount code the agent emailed that shopper; the exact checkout they abandoned; their email address; their phone number; and, as a last resort for a guest who used a different email and no code, their name together with their shipping address. Where more than one of their carts could match, the cart that was actually called always wins. That's what saves the credit when a shopper buys through the emailed link (which starts a brand-new checkout) or checks out under a different email entirely. Each recovery records how it was matched, and the Dashboard's recovered revenue counts only the ones a call drove. If the order is later canceled, the attribution is reversed.

Email pickup@exil.ai any time. We're happy to help with setup, billing, or anything this assistant can't answer.