Summary

Project: Create chatbot prototype for fictional company Cherie-Me Florals

Role: Conversation Designer

Responsibilities: Create bot personality, define intents and slots, organize utterances, design conversation flows, prototype conversation

Date: October 2021

My design process

🔹Plan user journey

 Cherie-Me* Florals is a fictional company created in response to the current trend towards wellness and self-care. The business tagline is, ***Because today is a special occasion**,* and it encourages customers

to celebrate and find joy in every day of life, despite challenges. The user journey is basically linear, with the bot collecting information from the customer to personalize a floral arrangement.

*Chérie-me is a French saying I heard used by one of my favorite makeup artists, Violette Serrat, to describe a simple act of self-care.

🔹Create bot personality

  *What type of language does the Cherie-Me Florals bot use, and avoid, to make customers feel like every day is a special occasion?*

Five emotions came to mind when I was planning the Cherie-Me Florals chatbot personality: appreciation, joy, celebration, nostalgia and love. Company values that influence the bot are abundance, quality, sharing, sustainability and growth.

Below is the Cherie-Me Florals bot personality charted on how it displays "The Big 5" personality traits. The bot is open, flexible, welcoming, cooperative, polite and detail-oriented. It wants to get the details of your order right!

Cherie-Me Florals bot personality plotted on this UXWC template based on "The Big 5" personality traits

Cherie-Me Florals bot personality plotted on this UXWC template based on "The Big 5" personality traits

🔹Create intents library

 *For an actual business, I would start with discovery and a content audit to research possible utterances.* 

Please see the image below for a sample file of intents involved in ordering flowers. An AWS a rule of thumb is to have as many as 20 utterances per intent, so this is just a snapshot illustrating the process

of creating training data, in other words, organizing a library of the many possible ways to make a single request.

 Note: I am not a developer and do not have experience with Amazon Lambda. However I understand that a **.csv** file of intents can be uploaded to the bot as a Zip **JSON** file, so that intents do not need to

be added manually. Use of the .csv format allows the freedom and flexibility to organize intents in Microsoft Excel.

A list of AWS Lex Intents I created as a sample. BusinessHours was suggested in the template

A list of AWS Lex Intents I created as a sample. BusinessHours was suggested in the template