Help

Insights Conversational Trends

Airgentic Help

The Conversational Trends tab on the Insights screen helps you understand what your customers are asking about over time — organised into stable themes you can track week to week.

While the other Insights tabs show individual conversations and point-in-time metrics, Conversational Trends reveals patterns: which topics are growing, which are declining, and what new questions are emerging.


How It Works

Conversational Trends uses AI to analyse the questions users ask your service. It groups similar questions together based on their meaning and organises them into a two-layer structure:

Questions → Clusters → Themes

Questions

At the lowest level, the system captures the standalone question from each user interaction — a clear, self-contained version of what the user was asking. Questions that are semantically similar (asking about the same thing in different words) are grouped together.

Clusters

Clusters are groups of very similar questions discovered during a specific analysis period (typically a week). The system uses AI to identify questions that share the same intent and groups them together.

For example, these questions would likely end up in the same cluster:

  • "How do I reset my password?"
  • "I forgot my password, how do I get a new one?"
  • "Password reset not working"

Clusters are ephemeral — each analysis run discovers fresh clusters based on the questions in that time period. This means clusters can reveal new, specific issues as they emerge.

Themes

Themes are stable, persistent categories that exist across time. While clusters come and go with each analysis, themes provide continuity for tracking and reporting.

Examples of themes might include:

  • Connectivity Issues
  • Password Reset
  • Returns & Refunds
  • Product Setup

Each theme has an AI-generated label and description that captures the common intent behind the questions it contains. When new clusters are discovered, the system automatically maps them to existing themes — or creates new themes if the cluster represents something genuinely new.

Themes are designed to stay consistent so you can track trends week over week and month over month.

How Clusters Map to Themes

Each analysis run:

  1. Discovers clusters of similar questions from the selected time period
  2. Uses AI to match each cluster to the most appropriate existing theme
  3. Creates new themes when clusters don't fit existing categories

This two-layer approach gives you both stability (themes for consistent reporting) and discovery (clusters for spotting emerging patterns).


The Trends tab shows your themes and clusters in three connected views.

Theme List

The default view displays a table of all themes for your service:

Column Description
Theme The name of the theme (e.g. "Connectivity Issues").
Status A badge showing the theme's current state — Active, Emerging, Declining, or Inactive.
Usage The total number of questions mapped to this theme in the most recent analysis period.
Trend An indicator showing whether usage is going up, down, or staying flat compared to the previous period.
First Seen When this theme was first identified.

Click View on any row to drill into that theme.

Theme Detail

When you select a theme, you'll see:

  • Usage Over Time — A line chart showing how question volume for this theme has changed across recent periods.
  • Clusters — A table of the specific question clusters from the latest analysis that belong to this theme.

Each cluster row shows:

Column Description
Cluster A short label describing the cluster (e.g. "Wi-Fi drops after firmware update").
Usage The total number of times questions in this cluster were asked. If 10 users asked similar questions, usage = 10.
Members The count of distinct question phrasings in the cluster. If those 10 users asked 3 different variations of the same question, members = 3.
Match A confidence score (0–1) showing how well this cluster fits the theme. A score of 1.0 means the AI is very confident this cluster belongs here.

Usage vs Members: A cluster about "password reset" might have 50 usage (50 total questions asked) but only 8 members (8 unique phrasings). High usage with low members suggests a common, well-understood topic. High members with low usage suggests many variations of a less common question.

Click View on any cluster row to see the actual questions.

Cluster Detail

The cluster detail view shows:

  • Example Questions — A sample of representative questions from the cluster (up to 5). These are chosen because they best represent the cluster's core topic. When you have more data, you'll still see a sample rather than all questions — the examples shown are the most representative ones.
  • Theme Mapping — Details about how this cluster connects to its theme.

The Mapping section shows:

Field Description
Theme The canonical theme this cluster is mapped to.
Method How the mapping was made — llm means AI matched it automatically, seed means it was created when the theme was first established.
Confidence How confident the AI was in the match (0–1). Higher is better.
Review Whether the mapping has been verified — suggested means automatically assigned, confirmed means a human has verified it's correct.

Use this view to understand exactly what users are asking and whether the questions are being grouped correctly. If a cluster seems mis-matched to its theme, the example questions will help you understand what's really being asked.


Understanding Theme Status

Themes move through a lifecycle as your service evolves:

Status Meaning
Active This theme has been seen consistently and represents an established category.
Emerging This theme is new or growing — it has appeared recently or is seeing increased activity. Watch these closely.
Declining This theme has seen reduced activity recently. Users may be asking about this topic less often.
Inactive This theme hasn't been seen in several analysis periods and may no longer be relevant.

The system automatically updates theme statuses based on usage trends across analysis runs.

How New Themes Appear

New themes can appear in two ways:

  1. Seeding — When you run your first analysis, the system creates initial themes based on the clusters it discovers. These become your baseline.

  2. Emergence — In subsequent runs, if a cluster is discovered that doesn't match any existing theme well, the system creates a new theme for it. These start with "Emerging" status.

Emerging themes are worth watching closely — they can indicate new product issues, changing customer needs, or gaps in your content.


Generating New Analysis

The system analyses questions periodically. To run a new analysis:

  1. Select a date range using the date picker at the top of the screen.
  2. Click Generate Clusters.

The analysis will process questions from the selected period, group them into clusters, and map those clusters to themes. This may take a few moments depending on the volume of data.

Click Refresh to reload the results after generation completes.


Conversational Trends is designed to help you improve your AI service over time. Here are some practical ways to use it:

Identify Content Gaps

If a theme or cluster shows high volume but your AI is struggling to answer (check the main Insights tab for completeness), you may need to add or improve content on that topic.

Track the Impact of Changes

After updating prompts, adding content, or making configuration changes, use Trends to see whether the affected themes show improved outcomes in subsequent periods.

Spot Emerging Issues

Watch for new "Emerging" themes or clusters that suddenly appear with high volume — these often indicate a new product issue, a marketing campaign driving questions, or a gap in your knowledge base.

Report to Stakeholders

The stable theme structure makes it easy to report on customer intent over time. Themes stay consistent, so your reports can compare apples to apples.


Tips

  • Weekly rhythm — Running a weekly analysis gives you a good balance between responsiveness and stability.
  • Focus on high-usage themes — Start by drilling into your top themes by usage to understand what's driving the most volume.
  • Review emerging themes — New themes may indicate genuine new topics, or they may be duplicates of existing themes that need to be merged (a future feature).
  • Check cluster examples — If a cluster seems mis-labelled, look at the example questions to understand what's really being asked.

← Back to Insights overview

You have unsaved changes