Can AI Answer Product Questions, Check Availability, and Take Orders on WhatsApp?

The job usually reads like this: "We need AI to answer product questions and check availability on WhatsApp, take simple orders, and tell customers when the item they wanted is back."

Three of those four things work well today with a knowledge-grounded WhatsApp AI agent. One of them — live inventory sync — has an honest boundary you should know before buying anything. Here's the breakdown, using WhatsApp AI Pro as the concrete example.

Product and Availability Questions: Answered From Your Catalog

The agent answers from company information you give it — and loading it takes about 10 minutes: chat the facts in, upload a file, or paste your website and let it extract the catalog, prices, and policies (the AI Replies tutorial shows all three paths).

From then on, "do you have the red one?", "what's the price for 200 units?", "do you ship to Brazil?" get answered from your data — in the customer's language, 90+ supported, at 2 a.m. Every reply passes an accuracy check against the knowledge base before sending, which is what stops the classic AI failure of inventing a price or a shipping promise.

The boundary that matters: availability answers come from the catalog you uploaded, not from a live query against your store's database. If stock changed an hour ago and you haven't updated the company info, the AI doesn't know. For most catalog businesses — where "availability" means we make this, in these variants, at this MOQ — that's exactly right. For a fast-moving retail store where stock flips hourly, it isn't; see the last section.

Recommendations From the Conversation, Not a Catalog Dump

Because the agent reads the whole thread, "suggest something for a beach wedding" or "what goes with the model X I bought last month" gets an answer based on what the customer already said — their sizes, past questions, budget hints — matched against your catalog. It's the difference between a salesperson who remembers the conversation and a link to your product grid.

Taking a Simple Order

The agent can carry an order conversation end to end: confirm the variant, quantity, delivery address, and preferred terms, and assemble that into a clear summary. What happens next is a handoff by design — the order lands with a human (via the morning queue, Team Mode, or the Enterprise Inbox) who confirms payment and enters it into whatever system you actually invoice from.

It does not process payments, generate payment links, or write into an order management system. For a $30 impulse purchase with in-chat checkout, a Shopify-integrated cloud bot is the better shape. For orders where a human confirms anyway — wholesale, custom, B2B — the AI doing the conversation and a person doing the commit is the workflow most teams want.

"Tell Me When It's Back": Restock Notifications

This works today as a win-back broadcast: when the item is back, you pick the conversations that asked about it, tell the AI why you're reaching out, and it writes a personal message to each customer — referencing what they actually asked — while skipping anyone it judges shouldn't be disturbed. The broadcast tutorial covers the safety rails.

Note the shape: it's you triggering the broadcast when stock arrives, not an automated watcher firing on an inventory webhook. Manual trigger, AI-written personal messages. On a regular WhatsApp number that's also what keeps you out of spam trouble — these are replies to people who asked, not a blast.

What This Setup Doesn't Do

No live inventory/ERP/Shopify sync. The knowledge base is what you upload; update it when stock materially changes. If you need hourly-accurate stock answers wired to a store backend, look at API-based tools with commerce integrations — Interakt (deep Shopify/catalog integration) or Tidio (e-commerce live chat) are the closest fits, both on the WhatsApp Business API with per-conversation fees.
No in-chat checkout. No payment collection, no payment links, no order writes into external systems. A human commits the order.
A desktop app runs it. macOS/Windows, connected to your existing number by QR code — keep a machine on for 24/7 coverage.

Frequently Asked Questions

How does the AI know my products and prices?

You load company info once — by chatting it in, uploading files, or pasting your website URL for automatic extraction. The agent answers only from that, and each reply is checked against it before sending.

Can it really check stock?

It answers availability from the catalog you uploaded. That's accurate as long as you keep the info current, and wrong if you don't. It does not query your store's database live — that's the honest limit, and the reason fast-turnover retail should use a Shopify-integrated API bot instead.

Can customers complete a purchase inside the chat?

No. The AI confirms variant, quantity, and address, then hands the assembled order to a person who takes payment and enters it in your system. There is no in-chat payment step.

Can it notify customers automatically when stock arrives?

Semi-automatically. You trigger a win-back broadcast when the item is back; the AI writes an individual message to each customer who asked, and skips people it shouldn't disturb. There's no automatic inventory-webhook trigger.

Does it work in my customers' languages?

Yes — it detects the customer's language and answers in it, 90+ languages, including transcribing and translating voice notes. Same knowledge base, no per-language setup.

What does it cost compared to API-based commerce bots?

$49/month flat (or $499/year), no per-message or per-contact fees, on your existing WhatsApp number. API-based commerce platforms charge a platform fee plus Meta's per-conversation fees — see our cost calculator to model a real month.

Try it on your own catalog: download WhatsApp AI Pro, paste your website into the knowledge base, and ask it your ten most common product questions in Suggest mode. $49/month flat, 7-day money-back guarantee.

Related guides: