Solo research project
User research, user experience, conversational UI
August 2020- October 2020
1 Design Lead, 1 Product manager, 4 engineers, 2 data analysts
UX Designer
User research, Interaction design, Visual design
2021
Exeevo is a fast-growing company with a mission to help businesses achieve their sales targets through innovative technology solutions. As part of their growth strategy, Exeevo identified the need for an intelligent assistant to help field teams manage their administrative tasks more efficiently and improve the overall customer experience.
In this case study, we'll explore how Gini helps field teams like ours navigate through their busy days more efficiently and improve the overall customer experience.
Let’s imagine a day in the life of a sales representative:
it's a busy day for a field team at Exeevo, racing to meet their sales targets while juggling a long list of administrative tasks. As a representative, you are always out there to achieve your target, and With so much on their plate, it's easy to get overwhelmed and make mistakes. This could add on to daily stress, missed opportunities, or disorganization.
For example, you need to find a meeting scheduled while you’re planning your daily/weekly work or you want to quickly make a note during a meeting or just add a reminder on the go, while tracking all of their KPIs. This is exactly when you might need an assistant to help you guide through such times. This is where we imagined an AI-powered personal assistant, integrated into the Exeevo app to step-in.
The goal is to design an interactive personal assistant, Gini, to help the sales reps plan their daily tasks and navigate day-to-day challenges more efficiently. It will help sales rep stay on top of their schedules by providing reminders, scheduling meetings for them, record notes, and access relevant customer information on the go. By automating time-consuming tasks and streamlining workflows, Gini helps teams improve productivity and accuracy, leading to better customer service and overall business success.
Given the sensitive nature of user data, protecting it is a top priority for Exeevo. As a result, there were constraints around accessing certain data required to train the assistant. These constraints meant that some features had to be developed with limited data. They were balanced with data-driven insights along with user privacy.
1. Boost field productivity
2. Increase ROI
3. Engage customers
4. Transition to hybrid work model
Through user interviews with a diverse range of sales users, I discovered some pain points in the users daily schedule. Like -
This helped establish some basic goals for the assistant.
Some basic goals for the assistant based on the discussions with the representatives -
Help them plan their day more easily leveraging assistant strategies with - calendar review & suggestions, important reminders and notifications, show work KPIs, etc.
An assistant who will listen and maintain the bullet tasks during the meetings by recording meetings & can pick up important keywords for next meeting.
Show details of their app shortcuts and other daily important tasks in one place like - apply for time off, capture notes, sync activity, etc.
To understand user personalities, here is a brief summary of users, their behaviors, goals, frustration and limitations.
This is to understand the tone of responses based on the chatbot's personality -
Warm - friendly, relaxed, witty
User - Will you marry me Gini?
Gini - Sorry, I’m taken 🙂
Professional - diligent, effective, insightful
User - Please set up my meeting.
Gini - The best time to call this customer is between 12pm and 2pm. Should I set it up?
Objective - straight to the point, simple, decisive, clear
User - Gini, are you a feminist
Gini - I believe in gender equality
Focussing on some sample use cases. Keeping them simple and to-the point for now:
Now that personality, tone and sample interactions are sorted, I am considering Scheduling a meeting as the key use case for the sample conversation dialog between Gini and the rep. It would establish how the rep would achieve their goal and along the way I figured some other scenarios that could come up during the conversation.
START
💻 Gini : Hi, I’m Gini. How can I assist you today? Check schedule, Reminders, New class
👩🏻 User : Can you schedule my meeting with Dr. Henry for tomorrow?
💻 Gini : Sure. What time should I schedule it? 11am or 12pm
👩🏻 User : 12pm.
💻 Gini : Done. 12pm meeting with Dr. Henry is added to your calendar tomorrow. Is there anything else I can help you with - Schedule, Meetings, Reminders?
👩🏻 User: What are my other meetings for tomorrow?
💻 Gini : You have an Allotment pick-up at City Hospital by 9am and a meeting with Dr. Chad at 10am. Do you want me to add anything else on tomorrow’s schedule? - Reminders, New meeting
👩🏻 User: Remind me to send a follow-up mail to Apollo Pharmacy coordinator by 4pm tomorrow.
💻 Gini : Done, I will add that to reminders. Anything else? - New meeting
👩🏻 User : No, thanks.
END
Diving a little more into details with a flow diagram. It would help to give an overall view of how the conversation should flow to help the user achieve his goal. This will help to figure out along the way what are other probable scenarios that could come up and how to handle those and redirect the user back to the conversation. Symbols used in the flow map -
This helped in streamlining the design and development workflow, and acts as a single source of truth for a scalable UI and streamlined UX guidelines.
Symbols used in the flow map -
Flow diagram for the conversation & sample scenarios -
Once the flow is finalized, I decided to mock how the conversation would look visually in a mobile app. I refined a few sample dialogs in this design stage & designed a few solutions.
According to recent study, 57% of businesses agree that chatbots deliver large returns on investment (RoI). The bots can capture and process customer sentiments from conversations, without explicitly asking, using the sentiment analysis approach. It can help businesses understand whether users are responding positively or negatively and make their bot more user friendly.
As a designer, I believe that in any product there's always a lot of scope to improve the user's experience. After more research, I found some add-on improvements on the existing bot model that could be considered in the future -
