A chatbot for your online store: what it should answer
Updated: · 6 min read
In short
A chatbot for your online store should answer the questions shoppers ask before and after buying: shipping, returns, warranty, payment, sizes and availability. Feed it your real policies and product pages, review what it learned, and let it hand complex cases to a person. It answers from stored information, so live order status needs a developer integration.
A chatbot for your online store works best when you decide in advance what it is there to answer. Shoppers are not chatting for fun. They want to know if a product fits, when it arrives and what happens if they change their mind. This guide covers the questions an ecommerce chatbot should handle, what content to give it, where its limits are, and how to measure whether it helps.
Why online shops get so many repeat questions
Most stores sell the same way every day, so the same questions come back every day. How much is shipping? Can I return this? Does it come with a warranty? Each one takes a minute to answer, and a minute many times a day adds up.
Shoppers also ask at awkward moments, such as at night or during a sale, when your team is asleep or overloaded. A shopping support bot gives them an answer when they are still deciding, which is when it matters most.
The questions your ecommerce chatbot should cover
Start with the topics that fill your inbox. This table is a practical starting list; adapt it to your store.
| Topic | Typical questions | Where the answer lives |
|---|---|---|
| Shipping | Cost, delivery time, areas you serve | Shipping policy page |
| Returns and refunds | Time limit, condition, who pays | Returns policy |
| Warranty | Length, what is covered, how to claim | Warranty terms |
| Payment | Accepted methods, installments, invoices | Checkout and FAQ pages |
| Product details | Size, materials, compatibility, care | Product pages and spec sheets |
| Contact and hours | Phone, support hours, location | Contact page |
What content to feed the bot
The bot is only as accurate as the information behind it. Ravano can learn from your website address, files in PDF, Word, Excel or CSV, a Telegram channel, an Instagram page, a database or text you paste in.
Policies first
Shipping, returns, warranty and payment rules are the highest-value content. Make sure each is clear and complete in one place. If your policy is vague, the answer will be vague too.
Product information
Add product pages, size charts and spec sheets. A spreadsheet export of your catalogue with names, descriptions and key attributes can work well. Remove discontinued items so the bot does not recommend them.
Review before launch
After the bot reads your sources, you review five sections of what it learned and approve them. Check prices, deadlines and conditions carefully. If two pages disagree, fix the source.
Order status: what a bot can and cannot do
This is the question shop owners ask most, so here is the honest answer. A chatbot built on stored information can explain how order tracking works: where to find the tracking number, how long processing usually takes, and who to contact if a parcel is late.
It cannot tell a specific customer where their specific parcel is. Ravano answers from your stored, approved information and does not see live order data. If you want that, a developer can integrate your order system through the REST API.
Until then, write the order-status answer so it is useful on its own, and set the bot to hand over to a person for questions about an individual order.
Set a tone that fits your brand
A fashion boutique can be warm and playful. A parts supplier should probably be brief and exact. Choose a personality on purpose, and read a few test answers to be sure it sounds like your store.
Keep replies short. Shoppers on a phone do not want an essay. Give the answer, add the source, and invite a follow-up.
Know when to hand off to a person
Some conversations should not stay with a bot. Define them before launch so the bot knows when to step aside.
When the bot has no approved answer, it says so and offers a human instead of guessing. That is better than a confident wrong answer about someone's money.
- Complaints and angry messages.
- Damaged, wrong or missing items.
- Refund exceptions and anything outside the written policy.
- Wholesale quotes and custom pricing.
- Questions about one specific order, unless you have integrated order data.
Prepare for sales and seasonal peaks
Sales change the answers. Shipping may be slower, returns windows may differ, and discount rules may apply to some items only. Update your sources before the campaign starts, review the change, and approve it.
After the sale, update again. An outdated bot that still promises last month's offer causes more work than no bot at all. Treat the sources as a living document, not a one-time upload.
How to measure whether it works
Keep measurement simple and honest. In the Ravano panel you can see conversations, how many answers came with sources, how many were handed to you, and the questions the bot could not answer.
For the wider picture of how bots and people share the work, see chatbot vs human support, and for the setup basics read how to build a chatbot for your website.
- Read a sample of real conversations each week.
- Look at the unanswered questions. Each is missing content you can add.
- Watch hand-offs. If one topic keeps going to humans, improve that content or accept that it needs a person.
- Ask your support team which questions they now see less often.
Launch your online store chatbot
A good online store chatbot is mostly a matter of clear policies, accurate product information and sensible hand-off rules. Ravano handles the rest: it reads your sources, lets you approve them, and installs on your shop with one line of code.
Building and testing are free. When customers start using the bot, you pay 650 Toman per request from your wallet; see pricing. Create your free account to test it with your own policies, or look at the chatbot for online store page for more.
Frequently asked questions
What should an ecommerce chatbot answer?
Start with shipping, returns, refunds, warranty, payment methods, product details and contact hours. These are the repeat questions shoppers ask before and after buying, and the answers already exist in your policies and product pages.
Can a chatbot tell customers where their order is?
Not by default. Ravano answers from the information you stored and approved, and it cannot see live order data. A developer can connect your order system through the API, otherwise the bot explains tracking and hands over to a person.
How do I add my products to the chatbot?
Give it your website address, or upload product sheets as PDF, Word, Excel or CSV files. You can also paste text. Then review what the bot learned and approve it, removing discontinued items so they are not suggested.
Will the chatbot give wrong shipping or return information?
It answers only from information you approved and shows the source, so errors usually trace back to an outdated policy. Review the learned content before launch and update your sources whenever policies or sale terms change.
How much does a chatbot for an online store cost?
Building and testing is free. Once real customers use it, each request costs 650 Toman from a wallet with a 100,000 Toman minimum top-up. The website widget is a one-time 1,490,000 Toman. There is no monthly subscription.
