How do I build a Text Analytics-based Insight Topic - evaluagent CX Smart Intelligence
Navigation prompt
Go to Analytics > Click Manage topics
By following the navigation prompt, you'll be directed to the Manage topics screen where you can see all topics available to your organisation and those provided by evaluagent.
To create a new Text Analytics-based insight topic, open the topic builder by clicking on the +New topic button.
Choosing Text Analytics
From here, options for technology are presented, allowing you to choose between Text Analytics and GenAI:
Text Analytics: Uses evaluagent's rules engine to determine whether queries (words and phrases) are present or absent in a conversation.
GenAI: Uses a large language model (AI) to judge whether or not context or intent is present or absent in a conversation.
Select Text Analytics.
Then, options for the type of topic are presented, allowing you to choose between Line Item and Insight topic:
Line Item: Line Items are added to a scorecard and then run during the evaluation process, presenting the results to the evaluator for review.
Insight topic: Insight topics are run across all conversations as they are imported, and the results are immediately available to you to filter and organise conversations before evaluating.
Select Insight topic.
The insight topic builder will slide into view, ready for you to begin populating it with details.
Configuring your Text Analytics Insight Topic
Name: You will use the name to identify the topic in your list of topics. Keep it brief and descriptive.
Description: Expand the purpose of the topic in the description. Users can read this to gain an understanding of what the topic is designed to detect.
Theme: The theme is used to group data on reports. Use the same theme name if you want all related topics displayed together on the report. Alternatively, find a theme that best fits where you want it on the report.
Configure queries
The configure queries section defines how the topic will detect patterns when analysing a conversation. You will define the words and phrases you want evaluagent's rules engine to seek out in the conversation and return results.
Adding query terms
You can add query terms in two ways:
Manual entry - Type words or phrases in the input field and press Enter or click Add. Use the dropdown to choose between creating separate queries for each comma-separated term, or treating the entire input as one query.
Bulk upload - Upload a CSV file with terms in the first column. Terms exceeding the 100-word limit are not added.
Configuring individual query terms
Once you've added your terms, you can configure each query 'pill' (word/phrase) individually. Click on the pill to view and set conditions for your query. These can include:
Speaker - defines who said the word or phrase: the agent, customer or either.
Slop - the maximum number of words that can exist between the start and end of a phrase or sentence (this is not presented when singular words are used).
Searched in order - defines whether terms must appear in the order they are set out using the THEN operator or in any order using the AND operator throughout the whole conversation (this is not presented when singular words are used).
Search across utterances - groups utterances into one text string either spoken by the agent, customer or both, depending on the speaker conditions set.
Search entities - whether or not entities are to be included in the search.
Query profiler - This adjusts how the search is executed and what results are returned. We recommend using the 'unified' or 'default' drivers in most instances.
Query term components
Each query term consists of span components that define what to search for:
Term: Exact word or phrase matching
Wildcard: Pattern matching with * and ? characters
Fuzzy: Approximate matching with configurable similarity
Synonym: Matches words with similar meanings
Phonetic: Matches words that sound similar
You can combine these components using THEN (sequential order) or AND (any order) logic.
Apply default speaker to all terms: You can set all terms to the same speaker rather than clicking on the speaker icon next to each term.
Pass conditions: Define how many of the terms should return a positive result before the topic is considered a pass result.
Using "Build with AI" for Text Analytics Topics
The "Build with AI" feature helps you create text analytics topics by generating query logic from natural language descriptions. This feature uses GenAI technology to convert your description and examples into keyword/phrase matching queries.
Important: This feature generates text analytics queries (pattern matching), not AI/GenAI evaluation prompts. For AI-powered evaluation, create a Generated AI topic instead.
How to use Build with AI:
Start a new Text Analytics topic
Click the "Build with AI" button
Fill in the Interaction Format (calls, email, live chat, etc.)
Enter an Interaction Theme (minimum 5 characters)
Provide a detailed Interaction Description (minimum 100 characters)
Add at least 3 examples of text that demonstrates the topic
Click Start topic build to begin generation (can take up to 3 minutes)
Once complete, the system creates a draft topic with text analytics queries that you can review and refine before publishing.
Publishing your Topic
Once you save your topic, it will appear in the manage topic table under the tab 'Our Topics'.
NOTE:
Before conversations are tagged with insight topics, they have to be made active from the burger menu drop down in 'Manage Topic' table.
A topic can be deactivated from the same place only if it has not been attached / linked to a scorecard.
Related Help Guides: