Answering 'is it in stock?' on WhatsApp with an AI catalogue bot
A retail store can answer 'is it in stock?' instantly by combining a product knowledge base with a live Google Sheets or inventory lookup on WhatsApp. The customer sends a product name or photo, the bot matches it, checks current stock by size or variant, and offers to reserve it for pickup, all without a staff member touching the phone.
- A knowledge base handles product questions; a live sheet lookup handles stock questions — you need both.
- Reserve-for-pickup converts window-shoppers into store visits better than just saying 'yes, available'.
- Update your stock sheet at least twice a day, or the bot will confidently say the wrong thing.
- Let the bot escalate to a human when a product isn't in the catalogue at all.
- Make stock lookups branch-aware if you run more than one outlet.
The question every retail store gets a hundred times a day
'Is this in stock?', 'Do you have it in size 32?', 'Kitne ka hai?' — these messages arrive all day on a store's WhatsApp, usually from people who saw a post or walked past the shop earlier. Answering each one manually pulls staff away from the counter, and slow replies lose the sale to whichever store answers first.
The problem compounds during a sale or a festive weekend, when the same handful of trending items get asked about dozens of times an hour. A staff member typing the same 'yes, available in M and L' reply repeatedly is exactly the kind of repetitive work automation is built for, freeing them to actually help the customers standing in the store.
Two different lookups: catalogue and stock
These are separate problems. A knowledge base (product descriptions, pricing, materials, care instructions) answers general questions well through AI Reply. But stock levels change hourly, so they can't live in a static knowledge base, they need a live lookup against your actual inventory sheet or system, typically through a Google Sheets or HTTP/API request block.
Getting this split right matters: a bot that guesses stock from a knowledge base will eventually tell someone a sold-out item is available.
A simple way to think about it: the knowledge base answers 'what is this product like', while the live lookup answers 'can I actually buy it right now'. Confusing the two is the single most common mistake retailers make when they first automate WhatsApp stock enquiries.
The reserve-for-pickup flow, step by step
Once stock is confirmed, don't just say 'yes'. Offer to hold it: 'I can reserve this for you until 6 PM today, would you like me to?' This does two things: it commits the customer to visiting, and it tells your staff exactly what to keep aside.
| Step | What happens | Block used |
|---|---|---|
| 1 | Customer names or photographs a product | AI Reply matched against knowledge base |
| 2 | Size/variant asked, stock checked live | Google Sheets lookup + Condition |
| 3 | Offer to reserve, with expiry | Buttons + Set Variable (timestamp) |
| 4 | Auto-release if unconfirmed | Wait / Delay (2-4 hours) |
A reservation should have a short expiry, a few hours, after which it releases automatically so stock doesn't sit blocked for no-shows. Edge case: if two customers ask about the last piece within minutes of each other, the first confirmed reservation should mark it held immediately so the second enquiry sees it as unavailable rather than both being told 'yes, available'.
Building it on Convo360
- Upload your product catalogue (descriptions, sizes, prices) to the knowledge base so an AI agent can answer general questions.
- Connect your stock sheet via the Google Sheets integration, mapping product code to available quantity per variant.
- In the Flow Builder, use a Condition block after the AI identifies the product, to look up live stock for the requested size or colour.
- If in stock, offer a Buttons choice: 'Reserve for pickup' or 'Just checking'.
- On reserve, use Set Variable to log the hold with a timestamp, and a Wait / Delay block to auto-release after a few hours if unconfirmed.
- Send a store-address location pin and store timings automatically with the confirmation.
Handling what the bot doesn't know, and mistakes to avoid
If a product isn't in your catalogue at all, or the question is unusual ('do you do alterations?'), the flow should hand off to your team through live-chat rather than guessing. A bot that says 'let me check with the team' and hands over cleanly builds more trust than one that bluffs an answer.
It's worth reviewing these handoffs periodically. If the same 'unknown' question comes up repeatedly, that's a signal to add it to the knowledge base or the guided menu, so future customers get an instant answer instead of a handoff. The most common mistake here is leaving the stock sheet update to 'whenever someone remembers' — set a fixed twice-daily update habit (opening and after lunch rush, for instance) instead.
Multiple branches and shared stock
Retail chains with more than one outlet face an added wrinkle: a product might be in stock at one branch and sold out at another. The stock lookup should be branch-aware, either by asking the customer which branch is nearest or by detecting it from the WhatsApp number they messaged, so the answer reflects the store they'll actually visit rather than combined chain-wide stock that isn't useful to them.
A simple fix that avoids a lot of confusion: keep a separate stock column per branch in your Google Sheet rather than one combined total, and have the flow ask 'which store would you like to check?' before running the lookup if the customer hasn't specified.
What this costs, and measuring it
Catalogue and stock-lookup flows run on the Starter plan at ₹999/month (WhatsApp + Web chat, flow builder, knowledge base). If you want the AI to also handle open-ended product questions with a language model, you'll need an OpenAI, Claude or Gemini key of your own, Convo360 adds no markup on those tokens.
To measure whether the bot is actually helping, check Flow Logs for two numbers: how many stock enquiries end in a reservation, and how many fall through to a human handoff because the product wasn't found. A rising handoff rate usually means your knowledge base needs a catalogue refresh, not that the bot itself is failing.
Example conversation
Industry guide: AI agents for Retail & Stores — Catalog enquiries, offers & store-visit bookings.
Build your store's stock-enquiry bot
Start free in minutes — no code, no credit card. Bring your own AI keys and pay no per-message markup.
Frequently asked questions
Can the WhatsApp bot check real-time stock, not just a static catalogue?
Yes, by connecting a live source like Google Sheets or your inventory system through a lookup block, so the answer reflects current stock rather than a fixed product list.
What happens if two customers ask about the same last piece at once?
The reservation should be time-stamped and the flow should mark stock as held the moment the first customer confirms, so the second enquiry sees it as unavailable or already reserved.
Does this work with product photos, not just text?
Customers can send a photo along with a description and the bot can match it against your catalogue if product images are tagged in your knowledge base; exact image recognition depends on how the catalogue is set up.
How often should the stock sheet be updated?
At least twice a day for fast-moving categories, and immediately after any large sale, so the bot doesn't confidently confirm stock that's already sold. A short opening and after-lunch-rush habit is usually enough to keep answers reliable.
Can this route to a store WhatsApp number I already use for orders?
Yes, if it's already on the WhatsApp Business API, or you can keep using your existing app number alongside the bot through Coexistence, which runs the AI on the same number as your phone app without disrupting either.
Related reading
Available across India — see WhatsApp & AI calling automation by state and city.