The AI Assistant lets you explore by asking questions in natural language. It retrieves relevant information from the project records and uses that context to generate an answer.
Get started with a question
Open the AI Assistant from within your text analysis project.
You can select one of these suggestions or enter your own question.
For example, you might ask:
What are the main issues people are experiencing?
What is driving lower experience scores or negative feedback?
What themes are associated with negative feedback?
How do experiences differ between different groups?
The AI assistant works within the scope of the text analysis project. Questions directly related to the feedback in your dataset will generally produce the most useful responses.
Ask more focused questions
Use findings from other Journey Insights Hub features to guide your questions.
For example, if Drivers shows a concept associated with lower scores, you might ask what people are saying about that topic in lower-scoring experiences. If Sentiment identifies a negatively discussed theme, you might ask which concerns or experiences contribute to that result.
This creates a natural workflow between exploring the project visually and using the AI assistant to investigate what you find.
Focused questions are generally more useful than broad requests. When relevant, include:
The topic you want to investigate
The group or time period you want to examine
The comparison you want to make
The type of answer you need
For example:
"Summarize the most common account-access problems in low-scoring feedback and explain which customer groups mention them most often."
Continue the conversation
Refine your questions as you learn more about the data. You can begin broadly to identify an area of interest and then ask more specific follow-up questions.
For example:
What are the main account-access concerns?
Which concerns appear most often in low-scoring feedback?
How do these concerns differ between new and existing customers?
What records best illustrate the difference?
Follow-up questions should remain connected to the information available in the text analysis project.
Validate important responses
Use the AI assistant to accelerate exploration, then use Journey Insights Hub’s analytical features and source records to validate the conclusion.
For important findings:
Locate the relevant concepts
Check their Volume
Compare applicable groups or periods
Review Drivers and Sentiment where relevant
Examine the underlying records in Records
Confirm that the evidence supports the response