Conversational Analytics

This page explains the key steps and best practices for this feature in Wejot.

In traditional survey tools, analysis is often a disconnected final step. After collection, users export data to Excel, SPSS, or Python, then perform complex operations to get conclusions.

This creates common problems:

• Non-experts struggle to run professional statistics
• The workflow is tedious and costly to iterate
• Weak feedback loops between data and research intent
• No real “data → insight” closed loop inside the product

Conversational analytics aims to make analysis as natural as asking a question.

You do not need to know which statistical method to use or how to code. You only describe what you want to know.

Wejot connects natural language directly to the execution layer through Conversational Analytics.

In a full analysis run, the system will:

• Understand your analysis intent
• Choose appropriate statistical methods
• Generate and execute analysis code automatically
• Produce charts, tables, and plain-language explanations

The process is transparent. You focus on the question and the conclusion.

Conversational analytics covers multiple levels—from basic review to professional analysis.

Get an overall view

Give me an overview of this survey’s results.

Wejot will typically output:

• response count, completion rate, average time
• basic distributions and means per question
• overall patterns for key indicators

Compare groups

Compare satisfaction by gender and city. Are the differences significant?

Wejot will:

• split the population automatically
• select appropriate tests (e.g., chi-square, t-test)
• explain whether differences are significant and what they mean

You do not need to handle statistical details—only interpret the result.

Explore relationships between variables

Are these rating questions correlated? Which questions matter most?

Wejot can:

• compute correlations
• identify strong/weak relationships
• visualize variable structures

Scales and modeling

For more advanced research, you can ask directly:

These items form a scale. Calculate reliability for me.

Or:

Run a regression and tell me which factors drive satisfaction.

Wejot will run reliability analysis, build regression models, and explain results.

Open-ended text analysis

For text questions:

Summarize the main themes in the suggestion responses.

Wejot will:

• aggregate responses
• cluster common themes
• extract representative viewpoints

Behind the scenes, Wejot does three things:

1. Turn natural language into an analysis plan
2. Generate and execute statistical code automatically
3. Package results as readable charts and explanations

This means:

• New analysis needs do not require pre-built product features
• Different study scenarios can be composed flexibly
• Analytical capabilities can expand continuously

Conversational analytics is a great fit if you:

• are unsure which method to use
• want to validate a hypothesis quickly
• need to iterate analysis angles repeatedly
• prefer spending time interpreting outcomes rather than operating tools

Its value is not to replace statistical knowledge, but to make data usable for more people.