"Is this in stock?" Answering it forty times a day
The short answer
Upload what you sell and what it costs, once, and let the app answer from that. The important part is that it answers from your actual list rather than improvising, and that it says it will check rather than guessing when the item is not on the list.
The WhatsAssist team ·
If you sell anything with sizes, colours or a limited run, you already know which message arrives most. It is four words long, it arrives at eleven at night, and answering it late is the difference between a sale and a screenshot sent to a competitor.
Why this one message matters more than the volume suggests
A stock question is not a browsing question. Somebody asking whether the blue one is available in medium has already chosen the item, the colour and the size. They are one sentence from buying and one silence from leaving.
That is why the reply time on this specific message is worth more than the reply time on anything else in your inbox. A vague answer costs you the same as a slow one: “let me check and get back to you” at 11pm means the sale happens tomorrow if it happens at all.
And it repeats. The same item, asked about by different people, forty times between Thursday and Sunday. Nothing about answering it the fortieth time is more interesting than the first, which is exactly the profile of work worth handing to software.
What the app needs to answer it properly
The difference between a useful reply and a plausible one is entirely in what the app was given.
Your list. Products, variants, prices. A PDF, a photograph of a printed list, or text pasted in. What matters is that it comes from your actual inventory rather than from the model’s imagination.
What you are willing to state outright. There is a real difference between an app that says “yes, the blue is in, in medium” and one that says “we should have that, let me confirm”. The first is only acceptable if it is reading something you gave it and stand behind. On our paid plans that is exactly what grounding means: the verified details are handed over as facts the reply may assert, and everything else is hedged or handed to you.
Your past replies. A chat export teaches it how you actually write, which matters more in DMs than anywhere else. Sellers have a voice, and customers notice when it changes.
The two failure modes, and which one to prefer
| Says too much | Says too little | |
|---|---|---|
| Looks like | Confirms stock it does not know about | ”Let me check” for everything |
| Costs you | A cancelled order and an apology | The same sale, slower |
| Caused by | An app improvising with no inventory | An app with nothing uploaded |
| Fix | Upload the list, and only let it assert what you verified | Upload the list |
Both are fixed by the same thing, which is why the upload step is worth twenty minutes. Of the two, prefer the second: a customer who is told you will check stays a customer, and one who is told yes and then told no does not.
Instagram and WhatsApp are two conversations, not one
Worth knowing before you set this up: a customer who messages you on Instagram and later on WhatsApp is two separate conversations, in two separate apps, with two separate histories. An app that answers in both will answer each from its own thread.
Practically, that means the person who asked about the blue dress on Instagram in March and writes to you on WhatsApp in August is, to the software, a new person. Plan your own follow-up around that rather than assuming the tool stitches the two together.
What to upload before a drop
A launch is where this pays for itself, and it is also where a stale list does the most damage.
- The new items, with variants and prices, before the post goes up rather than after the messages start.
- What is sold out from the last drop. The most expensive wrong answer is confirming something you no longer have.
- Your shipping answers. Cost, timing, which countries or cities, what happens if it does not fit.
- Your returns policy, in the words you would use, not the words a template would.
Then send yourself the four questions you expect, from another phone, and read the answers as a buyer.
When it should stop
Not every DM should be answered by software, and a tool that does not know the difference is a liability rather than an asset. A customer chasing a late delivery, a complaint, a request for a discount, anything involving money moving the wrong way: those want you. The reply should stop and wait for you with the reason attached, and you should be able to see at a glance which ones it held back.
That behaviour, and the switch that turns automatic sending off entirely, are described on the features page. The plans, including which one carries grounding on your verified stock and prices, are on the pricing page.
The four messages that come after the stock question
Answering “is this in stock” quickly is worth little if the next four messages sit unanswered until morning. In practice a buying conversation runs a fixed route, and all four of the follow-ups are answerable from the same upload.
- “How much?” Straight from your list, including the variant price if the large costs more than the medium.
- “Do you deliver to me?” Your zones and your minimum. This is the question that most often ends a conversation when it goes unanswered.
- “How long?” Your normal dispatch answer, phrased the way you would phrase it rather than as a promise you did not make.
- “How do I pay?” The methods you accept, and nothing else. This is the point where a tool improvising becomes expensive.
If a tool answers the first question and stalls on these four, you have moved the bottleneck rather than removed it.
What to do about the item that is nearly gone
The hardest case in retail is the last one in stock, and it is worth deciding your policy before the software meets it. Two customers asking on a Saturday evening about the same remaining item will both be told it is available, because nothing in a document you uploaded knows one of them is about to buy it.
Sensible handling: keep the last unit off the list, or set your policy so anything close to sold out is answered as “I have one left, shall I hold it for you” rather than a flat yes. A held item is a sale; a double-sold item is a refund and a bad review. The tool cannot make that judgement for you, and any vendor implying otherwise has not run a shop.
One more thing about photographs
Customers send pictures. “Do you have this one?” with a screenshot of your own post is among the most common messages a seller gets, and it is worth knowing that an app reading notifications does not receive the picture. A photo arrives as a placeholder rather than an image, so a reply written from it is a reply written from nothing.
The practical handling is that these should reach you rather than be answered. It is a small category and a predictable one, and any tool that produces a confident answer to a message it could not see is telling you something about how it handles the rest.
Related reading
For a kitchen rather than a shop, the same argument runs through taking restaurant orders without losing your evening. And the honest account of what happens when an automated reply gets something wrong is here.