Kubbly guides
What conversational commerce needs to complete an order
Conversational commerce lets a customer ask questions and buy through a conversation. To complete an order or booking, the assistant needs more than a convincing answer: it needs the right business context, an enabled action in the connected system, and a result the customer and brand can verify.
Follow the transaction beyond the chat
A useful shopping conversation can start with a simple request: 'Can I get my usual?' The assistant still has to establish what 'usual' means, which store should fulfill it, whether the requested item is available, and what the customer wants to confirm. A fluent response does not establish any of those facts.
Separate three experiences during evaluation. An assistant may answer questions, direct the customer to checkout, or complete an enabled action in the merchant's system. All can be useful. They solve different parts of the journey, and buyers should be able to tell which one they are seeing.
Keep the business system authoritative
The product catalog, order system, and booking system determine what can actually be sold or reserved. The conversation should use the relevant store, product, and fulfillment context. When that information is unavailable or ambiguous, the assistant needs a defined next step rather than a confident guess.
For a booking, distinguish the customer's requested time from a confirmed reservation. For an order, distinguish a proposed basket from an accepted order. Ask the vendor to show the record created in the connected system and the status returned to the customer. Decide separately how payment is completed for the workflow you are enabling.
Make confirmation specific
Before a consequential action, the customer should understand what they are confirming. In an illustrative reorder, that means the items and quantities, store, fulfillment choice, and applicable total. A booking has a different confirmation: location, date, time, party size, and any conditions the guest needs to know.
Test what happens when the customer changes a detail before completion. Then test an interrupted conversation or repeated request. Ask how the system determines whether an order already exists before trying again. These checks reveal far more than a scripted conversation in which every answer is predictable.
Plan the handoff
Customers ask questions that do not fit the happy path. A missing item, uncertain availability, a complaint, or an action outside the assistant's permissions needs a clear route to a person. Define which requests the assistant can complete and what information the human receives when the conversation is transferred.
The customer should be able to understand the current state: still choosing, waiting for confirmation, completed, or needing help. A friendly tone cannot compensate for an unclear order status.
Connect outreach to the conversation
Conversational commerce also changes what happens after a campaign. A brand can use customer context to choose a relevant moment to reach out, then handle the reply as the start of a useful conversation. That requires coordination between customer data, campaign execution, and the assistant.
Kubbly connects customer context, campaign execution, and BrandRep conversations. With the relevant integrations enabled, its agents can complete orders in Shopify, WooCommerce, Toast, and Square, and reservations in OpenTable. Confirm the actual connection and actions for your brand before launch.
A practical evaluation checklist
- Ask to see a real connected-system record, not only a chat transcript.
- Confirm which channels and transaction actions are enabled for your account.
- Test an unavailable item or time, a changed request, and a repeated request.
- Agree on the customer confirmation and the route to human help.
- Measure completed transactions separately from messages, replies, and clicks.