I thought about what could be an easy way to call bots when you know exactly which one you need -different than the case you know what you want to do but not sure about which bot to use-
Now in Facebook Messenger you see a dedicated button on the keyboard (like the one with the transportation icon for Uber). But if there will be a lot bots, hard to fit them all on a keyboard. Opening a new screen for “More” bots during a conversation, then searching there for a bot isn’t good either.
As a solution, you could call them by mentioning just as you would do with your friends in FB comments or Twitter. Type “@” and either continue to type the name or select from the autocompleted suggestions. In this concept I added category titles in the auto-suggested list like People / Groups / Bots… So if say the bot you need doesn’t show up, you can search it directly under “Bots” instead of going through the long list of all kinds of results.
I tried to improve conversational commerce experience by combining Augmented Reality and video call, to make it feel like shopping together with friends in real world and getting their suggestions. It could also work for a “virtual try on” of eyewear, watches, jewelry or home furniture.
You could have the bots’ assistance while trying on, asking like “Show me the black color of this model” by using voice or type. You could also get personalized suggestions based on your face shape, which could be saved among your customer data for a more personalized shopping experience in future.
Also 360° views or videos could be used instead of just seeing a small plain image of the product.
Another idea here is to make multiple bots work together, passing the task to each other whenever needed, so that the user has a seamless experience. For example in case Uber Business is used for shipping, its bot can get the user’s location and show shipping details like price and delivery time on Checkout screen.
Today’s bots ask the user series of questions and let them type answers, which is annoying and time-consuming, whereas a bot should facilitate the task rather than making it more difficult. So I thought that widgets could appear in place of the keyboard when needed. Each task can have its specific UI elements, buttons or sliders when needed..
In this case, instead of asking the user a series of multiple-choice questions that narrow down their search by filters like gender, price etc., these selections can be made more quickly and easily on a GUI element.
The selections can be also prepopulated according to user data and previous choices. For example if the user calling the bot is a man, Gender would be selected as Men as default, which can be changed if needed…
