Customer Support That Answers Every Order, Return, and Sizing Question While You Sleep

A shopper has a full cart and one question: will this jacket fit if she normally wears a medium in another brand. It is 11pm. Your support inbox opens at 9am. By morning she has bought from someone who answered. That gap, the silence between a question and an answer, is where most online stores quietly bleed revenue. Not from bad products. From nobody being home.

I have spent eighteen years watching service businesses lose customers to slow replies, and ecommerce is the cruelest version of it because the store never closes but the support desk does. You sell at 3am. Your shoppers expect to be helped at 3am. Hiring a round-the-clock team to cover that is expensive and hard to staff. So the questions pile up overnight and you answer them when it is already too late.

The channels that actually matter for a store

Let me be honest about something. A lot of AI support pitches lead with phone calls because phones photograph well. For an online store, that is mostly noise. Your customers do not call. They open the chat bubble, they reply to the order confirmation email, they text back the shipping update.

So for ecommerce, the order of importance usually looks like this:

• Web chat is the front line. It is where the pre-purchase question lives, and a pre-purchase question that goes unanswered is an abandoned cart with a receipt.
• Email is where post-purchase lives. Where is my order, I want to return this, the size was wrong. Volume is high and most of it is repetitive.
• SMS is where shipping and order updates get replies, and where you can win back a hesitating cart.
• Phone still matters for higher-ticket items and the occasional angry escalation, but it is rarely the main event.

LastWorker handles all four, so you are not stitching together a chat tool, a help desk, and a texting app that none of which talk to each other. One agent, every channel, same answers.

Pre-purchase: the questions that decide the sale

Most of what kills a cart is a small, specific doubt. Does this run large or small. Is it in stock in my size. What is your return window if it does not work. When does it ship and will it arrive before the weekend.

These are not hard questions. They are just questions that need an answer in the next thirty seconds, not the next morning. The AI learns your sizing notes, your stock logic, your shipping cutoffs and your return policy in a setup conversation that takes about fifteen minutes. No code, no developer ticket. Then it answers in plain language, in the shopper's language, sub-second on chat.

Ninety-seven languages, by the way. If you sell internationally, you already know the support headache that comes with it. A shopper in Berlin asking about delivery to Germany gets an answer in German, instantly, without you hiring a German speaker.

Post-purchase: where repeat business is actually made

Here is the part owners underrate. The sale is not the relationship. The relationship is what happens when something goes slightly wrong and you make it right fast.

Most of post-purchase support is the same handful of conversations:

• Order status. Where is my package. The AI can answer with the tracking details and the expected date instead of making someone wait on a human to look it up.
• Returns and exchanges. Starting a return, explaining the window, swapping a size. This is the highest-volume, most repetitive ticket category for almost every store I have seen, and it is exactly the kind of thing that does not need a person.
• Wrong or damaged item. The AI captures the details, the order number, a photo if needed, and routes it to a human with everything attached so your team is not playing twenty questions.
• Sizing and product follow-ups. The medium was too tight, what do you recommend. Answered from your own product notes.

When something genuinely needs a human, a refund judgment call, an upset customer, a wholesale inquiry, it transfers or escalates with the full conversation history. Nobody starts over. The customer does not repeat themselves three times, which is the fastest way I know to turn a small problem into a one-star review.

Cart recovery without the creepiness

A shopper asks a question on chat, gets a good answer, and still drifts off. That is a warm lead, not a lost one. Because the AI handled the conversation, it knows what they were looking at and what they were unsure about. A well-timed follow-up over SMS or email, answering the actual doubt they had, recovers carts in a way a generic "you left something behind" blast never will. The difference is the message is about their question, not your discount code.

What it costs, and why the math works for stores

There is no monthly fee. You load a prepaid balance and pay per conversation: voice is $0.05 a minute, chat and SMS are priced per message, email is per resolved ticket. Set auto-reload if you do not want to think about it. A dedicated phone number, if you want one, is a dollar a month.

For a store, that pricing fits the way demand actually moves. Black Friday hits and your support volume triples in a weekend. You pay for the conversations you have, not for a fixed seat count you guessed at in January. In a slow week you pay less. You are not carrying overnight staffing cost for the two questions that come in at 4am, but those two questions still get answered.

Here is a rough sense of where each channel fits:

Why speed and 24/7 are the whole point

I will say the quiet thing plainly. The reason this works for ecommerce is not that AI is clever. It is that your store sells around the clock and your competitors are answering questions you are not. Speed is conversion. A question answered in seconds keeps a cart alive. A return handled cleanly turns a frustrated buyer into a repeat one. Coverage at 2am means the international shopper who would have bounced instead checks out.

You can compare the approach against alternatives on the pricing page or read how it stacks up on our comparisons. But the core idea is simple. Your store never closes. Now your support does not either.

If you want to see how it learns your catalog and policies, the setup is a fifteen-minute conversation, not a project. Most owners I talk to expect it to take a week. It does not, and that surprises them more than anything else about it.