Designing for the customer’s next step
Building BotAnswer.ai and CounterMint™ has made me pay closer attention to the work a product leaves with its customer.
A customer can get a correct answer and still be left with a difficult experience.
They know the appointment is available, but have to explain the request again to book it. They receive an offer for a collection, but cannot tell which number belongs to which item when they look at it later. They agree to something in a conversation, then have to chase the business to find out what happens next.
These are the moments I keep coming back to while building BotAnswer.ai and CounterMint™. The products serve different businesses. The design concern is closely related: how much work do we leave for the customer to reconstruct?
The quote after the customer leaves
CounterMint™ starts with the buying counter at a coin shop. The dealer has a collection to examine, market evidence to weigh and a buying policy to apply. The customer wants an offer they can understand and decide on.
It is tempting to treat the calculation as the whole product. Accurate arithmetic matters, but the experience continues after the dealer has produced a number.
The customer may take the offer home. They may want to sell some pieces and keep others. The market may move before they return. If the only record is a page of shorthand and a total, both sides have work to repeat before they can even resume the conversation.
That is why the customer quote is central to CounterMint™. Each item has a place in the offer. The shop’s identity and terms stay with it. The customer can reopen it, see its current state and take the actions the dealer has enabled. A printed quote can lead back to the same record through its Quote ID and PIN.
The dealer’s controls belong in that experience too. Repricing limits, counter-offer settings and acceptance windows make the offer workable for the business. The customer sees a usable next step because the underlying agreement has been defined carefully.
The conversation after the answer
BotAnswer.ai approaches the problem through LINE, where a message can be part of a booking, a purchase or an ongoing customer relationship.
An answer is useful. A completed task is often more useful. If someone asks about a service, the journey can let them choose it and book an available time. If they want a product, the conversation can continue through a visual menu and checkout.
I want those journeys to make advanced messaging capabilities practical for the business operating them. The operator should be able to configure what runs automatically, where approval is required and when a conversation needs a person. They should not need to understand every LINE API detail to make a clear customer journey.
The same thinking applies when automation stops. A staff member needs the context and an action they can take. Where approval or rejection is the actual task, an Inbox action with a customer message is more useful than another blank text box.
The work underneath a simple next step
My route into software engineering came through networking, automation and Python-based test development. Writing code to exercise a system teaches you to look across layers: what a component depends on, what state it is in and how a failure becomes visible. I was working on those problems when I began experimenting with early generative AI tools.
In product design, I ask similar questions about a handoff. What information needs to travel with it? Who owns the unresolved decision? What should the customer see while they are waiting? What happens if they come back tomorrow?
Those questions connect the interface to the application underneath it. A clear status needs a real state to represent. A reliable action needs defined permissions and an outcome. A conversation history is useful when the next person can continue the work from it.
What I want these products to make easier
I am not trying to remove judgment from the businesses using them. A dealer still evaluates the item and decides the offer. A LINE business still decides how it wants to serve customers and where its staff should take over.
I want the software to carry more of the surrounding work: retaining the context, applying the configured rules and making the next action clear.
That is the part of customer experience I find most useful to design for. After the first answer or the first offer, can the customer continue without starting over?