Five Minute Kitchen
Five Minute Kitchen is an Amazon Alexa app where Alexa will be users’ assistant in meal prepping, giving you directions on cooking, & to assist you in finding recipes tailored to your needs & goals.
One of the first things to begin designing voice user interfaces is creating user personas. This is to figure out the needs & goals of the users by gathering their demographics, vital statistics, & everyday activities. My participant I chose to do a user persona said that they want to meal prep because,
- She works out & it will be easier for her to manage their classes & it’ll be easier as well for her siblings to eat the same meals.
- She also wants quick & easy meals to make so it will easy on her schedule.
- She needs ideas on healthy easy to make recipes & needs to know what to buy from the store.
To keep in mind, the participant does not like meals that take about 25 mins or more to make.
After the user persona, I developed a System Persona for Alexa.
This helps us with how empathetic we want Alexa to be & how she would respond the user interactions.
I then create user stories based on the user personas to help me determine the functionality of Alexa & the interactions
Next are user flows for voice based off of the user stories. This helps me figure out the navigation for responses & interactions. The user flows show the interactions from the start to the end, from when a user doesn’t decide on the next interaction with the response, No or if they do with the response, Yes.
I would then go on with creating voice scripts. Creating customized prompt responses for utterances. Such as the welcome, goodbye & help message prompts. Cuisine types, exact cuisine types, no selection of a cuisine type, instructions, repeat, next steps, different recipes, etc. Here is an example –
After the voice scripts, I then gather participants to do usability testing for voice & chatbot. The tests were conducted at home, in person & virtually. I read aloud the prompts to the participants for them to respond to in order to test. The tests took approximately 20 minutes. I also set up an audio recording for me to review after as well for the in person & virtual tests. I gave them 5 tasks to complete while testing.
Task 1: Launch & begin Five-Minute Kitchen
Scenario: You are a new user to Five-Minute Kitchen & want to start cooking on Five-Minute Kitchen.
Task 2: Choose a meal type/cuisine
Scenario: You want to start cooking & decide on the type of meal to cook
Task 3: Get a recipe suggestion
Scenario: You can’t think of a recipe to make, so you ask Five-Minute Kitchen to recommend you something.
Task 4: Get a different recipe, confirm recipe & ask for ingredients
Scenario: You don’t like the recipe that was suggested or you changed your mind about a recipe you were going to make, so you ask for another recipe & want the ingredients to it so you know if you have everything.
Task 5: Get the recipe, cook the recipe given & repeat steps.
After completing the testing, the tests went smoothly & the results were mostly positive with little errors made. All the participants thought Five-Minute Kitchen was simple to use, even though there was a bit of confusion on how to begin with their response during the launch. They understood how to use the skill & there wasn’t much difficulty understanding the system. There may only be a few tweaks to make on the skill for improvement.
There were little issues that have been resolved.
Issue 1: Missing Utterance for Meal Specification
Suggested Change – Add more utterances.
Add ‘’Can you recommend me a recipe?’’
‘’Can you recommend me another recipe?’’
Users may not be able to get a proper response from the system if they ask a question & it is not listed as an utterance.
Issue 2: Overuse of words the user uses & the system doesn’t understand
Suggested Change – Users will have to use less words & be more clear & concise with what they are asking the system.
The system may not be able to understand what the user is asking or what they want if there are too many words.
Issue 3: When getting the recipe & asking for the system to repeat, the system repeats the whole cooking instructions over from the top.
Suggested Change – I can label the steps like Step 1, 2 & 3 so that way users can either choose which step they want to repeat & it will be easier for users to get steps repeated.
Users would have to listen to the instructions all over again from when they first started & it could take a while to get to the steps needed to be repeated.
Issue 4: Not having the option for users to ask what the ingredients are after getting meal recipe.
Suggested Change – adding utterances & responses & prompts for ingredients.
Users may ask for the ingredients before they start cooking. They may want to know if they have all the ingredients to make the meal.
Issue 5: Ingredients are listed all at once when there are too many.
Suggested Change – I can have the system list 4 ingredients first & the system can ask the user if they want them to continue on with the ingredients.
Users may not be able to keep up with the system if ingredients are listed all at once.
There weren’t many issues during the usability testing, but there are a few issues that I can resolve that were stated & they have been revised. The testing participants gave me feedback & it was very helpful information for me to make improvements to the skill. The updates I will be making will be adding additional utterances so that way users won’t run into any issues when asking the system for anything, adding more features to the skill such as the option for users to ask for the ingredients & have the system list them out & having the system label the steps for when users ask to repeat them. This will make everything seem more organized & it will be easier on the user using Five-Minute Kitchen. After I make these updates, I will do some additional testing on the prototype to make sure the experience runs great & smoothly for the user without having any issues.
After fixing these issues, I went for the last usability test, & went into AWS Lambda to customize the skill code for resolution.
After making these changes, my participants are happy with the new flow.
As a bilingual designer, I would love to start incorporating & working with language conversation designing as well so these digital products can be accessible & inclusive to all. That would also help with providing better user experience.
Thank you for reading!