The resulting chatbot is structured as follows:
Posted: Tue Dec 24, 2024 10:43 am
the famous supermarket chain Whole Foods, we proposed creating a chatbot that would offer recipes to application users according to their requests.
Based on the experience Initially, it was assumed that the smart assistant would be something like a catalog: the client selects a category of dish - soup, side dish, salad, and so on, then the main ingredient of the dish, after which he proceeds to view recipes.
During the process, it was decided to equip the bot with the turning employment leads into loyal customers functionality of adding all the necessary products for the selected recipe to the application basket to further stimulate purchases.
Since the range of products in the network is limited, food technologists created such a recipe database for the bot so that it mainly consisted of ingredients that can be bought in this network. If some components were not sold, they were indicated in the recipe as additional.
As for the selection of a dish, we decided to supplement the push-button catalog with the ability to search by ingredient name, dish category, cuisine and diet with the ability to search for recipes by a text request of a certain format.
Solution
The chatbot greets the caller and offers to learn more about what it can do
Those who have already used the bot skip this step, and newcomers receive a message with brief instructions.
The user then either goes to the recipe catalog or enters their request according to the instructions.
Depending on the specified parameters, the bot offers a different number of recipes, the preview of which with the name, photo, calorie content and cooking time can be scrolled
When clicking on a recipe the user likes, its full description opens, after reading it, the user can add the list of products for the recipe to the basket
Next, the bot offers to find another recipe or go to the cart to complete the purchase.
Based on the experience Initially, it was assumed that the smart assistant would be something like a catalog: the client selects a category of dish - soup, side dish, salad, and so on, then the main ingredient of the dish, after which he proceeds to view recipes.
During the process, it was decided to equip the bot with the turning employment leads into loyal customers functionality of adding all the necessary products for the selected recipe to the application basket to further stimulate purchases.
Since the range of products in the network is limited, food technologists created such a recipe database for the bot so that it mainly consisted of ingredients that can be bought in this network. If some components were not sold, they were indicated in the recipe as additional.
As for the selection of a dish, we decided to supplement the push-button catalog with the ability to search by ingredient name, dish category, cuisine and diet with the ability to search for recipes by a text request of a certain format.
Solution
The chatbot greets the caller and offers to learn more about what it can do
Those who have already used the bot skip this step, and newcomers receive a message with brief instructions.
The user then either goes to the recipe catalog or enters their request according to the instructions.
Depending on the specified parameters, the bot offers a different number of recipes, the preview of which with the name, photo, calorie content and cooking time can be scrolled
When clicking on a recipe the user likes, its full description opens, after reading it, the user can add the list of products for the recipe to the basket
Next, the bot offers to find another recipe or go to the cart to complete the purchase.