Where should you start?
The real workflow of the system
This is the core of the new system: AI is not just a chat responder — it is the link from a conversation to a verifiable business workflow.
1. The customer reaches out
Start at Inbox to read the conversation, check Customer 360, and see what this customer has ordered before.
2. AI helps interpret and search
AI only uses approved tools to search products, check stock, suggest alternatives, or open a restock queue.
3. Create the order and checkout
Once the information is complete, the system creates a real order and sends a checkout link tied directly to it.
4. Receive payment and verify the slip
Customers can send proof of payment, but confirmation always still requires a human to review and click Confirm.
5. Ship and close out
After payment, the team follows up in Shipping until it is complete, and reviews the overview in Dashboard or Reports.
How the restock queue brings a sale back
Good for shops where some items run out often, but the customer still has real buying intent. Ending the conversation too quickly loses both the sale and the demand data worth keeping.
Don't lose the conversation just because stock ran out
Instead of ending with just "out of stock," the system can invite the customer to opt in for a restock alert, so the sales opportunity does not disappear along with that message.
Turn demand you cannot fulfill yet into a follow-up queue
The shop can see who is waiting for what, and use that as a signal for restocking, prioritizing sales work, or measuring which out-of-stock item is costing the most revenue.
Help the sales team close the sale systematically
When stock arrives, the team can open the queue, review the messages, and reach out only to customers who already opted in — instead of manually digging through old chats.
A few concrete examples
Fashion shop: black dress, size M is out of stock
AI should offer a similar size or style first. If the customer still wants the original, the system invites them to opt in for a restock alert — once stock returns, sales can follow up immediately.
Grocery store: a bestselling promo item is temporarily out
Instead of losing the customer, the shop can capture who is waiting, then use that list to gauge how much to restock next round and reach out when the new batch arrives.
Beauty shop: one core item in a routine is missing
AI can still suggest what is available, but if the customer really wants that exact item, the system should keep a restock queue to close the sale once the full set is back in stock.
Conversations by shop archetype
This section shows that BMS doesn't reply with one script for every business — it adjusts its tone to the shop's context while still relying on real product and stock data from the backend.
Mini Mart / Grocery
Customer asks
Do you have Tom Yum instant noodles, 6-pack? I want 3 packs, and 2 bottles of 1.25L Coke.
AI should reply along these lines
Sure! Let me check what's actually available first. If everything is in stock, I'll summarize the order and total right away. If anything is out, the system should suggest a similar item or offer to notify you when it's back.
Shows that AI understands fast, multi-item orders and is ready to move straight into creating a real order.
Fashion & Apparel
Customer asks
Does this dress come in size M? If not, can you suggest something similar?
AI should reply along these lines
Sure, let me check size M for this style first. If it's out, I'll suggest a similar style that's still in stock, and if you'd rather wait for the original, I can set up a restock alert for you.
Shows that AI does more than say yes/no — it understands sizing, alternatives, and capturing the sale for later.
Beauty & Personal Care
Customer asks
I have sensitive, acne-prone skin. Do you have a gentle cleanser + moisturizer set?
AI should reply along these lines
Sure, let me pick products suited for sensitive skin from what the shop actually carries, then put together a short routine for you — and check whether the full set is in stock or if something needs a restock alert.
Shows that AI can respond in a consultative way while still grounding itself in the shop's real product catalog.
Gadgets & Accessories
Customer asks
Do you have a case for iPhone 15 Pro? And can you recommend a compatible screen protector and charging cable?
AI should reply along these lines
Sure, let me check the case for your exact model first, then recommend accessories that are actually compatible. If the model you want is out of stock, the system should offer alternatives or set up a restock alert for you.
Shows that AI understands compatibility and can cross-sell related products.
The main pages you will use
AI ground rules the help page must make clear
Stock, prices, orders, payments, and permissions must always come from a verifiable backend — never from something AI guessed at.
What AI can do
Search products
Check stock
Summarize information
Suggest alternatives
Prepare a proposal
What still needs a human to confirm
Confirm payment
Refund
Cancel order
Adjust stock
Merge customer
Questions you are likely to run into on the new system
A note about this mockup
This page is still a static mockup to refine the messaging and information architecture first — it is not yet wired to search, analytics, or dynamic help content.
If this direction looks good, the next round will build out a production version — search, role-based quick links, and deep links into per-menu guides.
Related shortcuts
Current status
The help page has been reworked into this new mockup
Use this to review the direction before expanding into a full help center.