Designing a collection of interaction patterns that provide the customer with a personalized, context-aware, intelligent chatbot (Einstein Assistant) experience while guiding the customer through a personalized product discovery process on the Salesforce Website.
About the Project
Figma
Miro
Google Meet
Google Docs
Tools
UX Researcher
UX Writer
UX Designer
Role
Full Document Link
Links
SALESFORCE
UX Research · UX Writing · UX Designer
How might we design a human-centered AI chatbot experience for small businesses using the Salesforce website for the first time with specific focus on product discovery?
Project Goal
Endless Scrolling
Ineffective Product Discovery
Pushing to Contact Sales Repeatedly
Too Robotic
Key Problems in Einstein Chatbot
Behavioural Questions - Clarifying Questions
Behavioural Questions - Clarifying Questions
Our findings from user tests suggested that people are more likely to interact with a chatbot which engages in a natural conversation and assists them in the process.
Behavioral Questions - Suggestive Prompts
How might we assist users in writing prompts to get the most relevant and accurate information?
Users expect a Gen Ai chatbot to understand their needs and be context-aware.
Comparing Products
How might we simplify product discovery and recalling past interactions to ease decision-making?
Users expect the chatbot to simplify product discovery by providing quick information to help them decide whether to explore further.
Timeline
Users expect an easier way to refer to relevant information in the chat, such as product suggestions and their inputs, reducing the time spent scrolling back and forth.
help users navigate through different sections within the lengthy chatbot conversation, and retrieve needed information more easily?
Users can keep track of chatbot conversation based on topics of entries. And jump to specific section to find relevant responses they need.
Behavioural Triggers - Connect to an agent
Users assume the conversation with a chatbot should provide necessary resources without contacting an agent.
Users get frustrated when pushed to connect with agents too often, preferring more to time to explore products first.
How might we introduce agent support without disrupting the self-service experience?
Behavioural Triggers - User Inactivity with the Chatbot
Salesforce values customer retention and aims to ensure users feel supported throughout the journey.
How might we re-engage users during inactivity?
Research & Discovery
To explore different types of responses/personality, we narrowed down to either Co-pilot or ChatGPT. While Salesforce's AI chatbot may not possess the extensive knowledge base of generative AI, we aimed to assess participants' responses to its personality and tone. Co-pilot and ChatGPT allow us to personalize the personality and tone. However, Co-pilot provided different answers every single time, and we wanted some consistency, especially since we have to test on multiple participants and analyze the different responses of the participants. Thus, we went with ChatGPT which provided consistent responses.
We input 3 different personality types; fun/joy personality, formal personality, and empathetic personality. We are looking to observe our participants on:
Which tone and personality do users appreciate or seem to like?
Which tone and personality welcomes new users better, and which users find more engaging?
Which tone and personality seems to communicate information better? Do tone and personality have an effect on the efficacy of communication?
Which tone and personality is appreciated in trouble-shooting/problem-solving scenarios?
Who do people imagine they’re “talking” to?
User Research with ChatGPT
Designing the chatbot’s personality and tone, basing them on how the customer seems to want to interact with the chatbot (with insights based on what web page the customer is currently visiting).
Goal
After conducting 16 user research, we identified several personality traits that users appreciate in a chatbot designed for professional environments. Users value a chatbot that is detail-oriented, offering clear and precise information while guiding them through processes with structured, step-by-step instructions. They prefer a chatbot that avoids jargon unless explicitly requested, making it accessible for all experience levels. The chatbot should be opinionated in a conclusive way, providing actionable insights only after understanding the specific needs of the business context. Users also appreciate a chatbot that proactively offers suggestions and directions, facilitating smooth workflows and next steps. A formal tone, free of emojis, is preferred to maintain professionalism. Additionally, users seek a chatbot that takes initiative by anticipating needs, shows sympathy for user challenges, and remains curious and communicative to ensure clarity. Being result-oriented with a focus on retention is essential, as this demonstrates a commitment to both immediate support and long-term user satisfaction.
Finalised Personality Traits Through User Research
See Full Project