Hardware components | ||||||
 |
| × | 1 | |||
Software apps and online services | ||||||
 |
| |||||
Gifter started as an idea around a need to find high-quality gifts for your friends and family. Today the process of gift searching is tedious. You have to get on the computer, go on various sites to find highly rated well-reviewed products and then compare them.
We thought this is the type of task that is perfect for an Alexa application. Why go through all the tedious searching when modern machine learning algorithms can do it for you. Simply ask Gifter to do it for you!
IMPORTANT:
Person submitting this project is not the owner of AWS account where skill lives. The owner is Vamshi Ponnapalli and he is a member of the team.
AWS AccountID: 184891664996
Skill ID: amzn1.ask.skill.f4198d76-27d1-4ad6-8201-f02048d98654
Description
In order for Gifter to work well, we gather price range and product categories user is interested in. For the price range, we restrict the user to numbers, while for categories the user can say anything from 'video games' to 'purple bear fur'. We do require that user provides at least two categories before they can continue. These pieces of information are gathered by our Gifter Skill interacting with the user in step (1). Look below for a full VUI diagram.
Once our skill has gathered all the information provide a recommendation it calls our 'Recommendation Engine' lambda. Ideally, that lambda will be able to store and maintain various offers from various internet providers and categories. For the purposes of completing this competition on time, we limited its behavior to going to searching for products on Amazon.com. We search by categories user requested and then pick out highest rated, best-reviewed products and arranging them in the order of popularity, steps (4), (5). Top 3 products are then returned to Gifter skill lambda, step (6).
Once Gifter skill has received our product recommendations it will send a card with the product name, picture, price, rating, and URL. The user is then asked if s/he would like to see another recommendation if so we send them a card with another product recommendation. These process can be seen in steps (7), (8) and (9).
VUI Behavior
Follow diagrams depict user interaction and voice UI behavior between the user and the Gifter skill.
Comments