Manage Categories

Required permission: Interaction Analytics Admin

You can modify categories Data groups that Interaction Analytics uses for parsed call transcript data. They make it easier to find trends. and category folders. You can merge, split, move, delete, and add new categories. You can also modify the category rules Criteria used to determine whether a transcript is placed in a certain category. Rules contain sets of keywords or phrases a transcript must include or exclude. that dictate which interactions The full conversation with an agent through a channel. For example, an interaction can be a voice call, email, chat, or social media conversation. fall into the category. You can pull in a sample set of interactions to see how the current categories are working. This helps you see how your changes affect the sample data before you apply those changes to a category set. If you make changes to categories applied to workspaces Named view of one or more widgets., those changes won't take effect until you click Update Data on the Language Model page.

Manage Categories

You can rename, move, duplicate, and delete categories Data groups that Interaction Analytics uses for parsed call transcript data. They make it easier to find trends. you've created.

1. Click the app selector and select Analytics.

2. Click Language Model > Categories.

3. Click more options in the row of the category you want to manage. Select the action that best suits your needs.

   Learn more about actions in this step

   Field	Details
   Rename	Enables you to rename the category.
   Move	Enables you to move a category to another folder or location. To rearrange the categories, you can also click and hold while dragging the category to where you want it.
   Duplicate	Enables you to copy a category. This is helpful if you want to create a category similar to an existing one with slight modifications.
   Delete	Deletes the category.

Manage Folders

1. Click the app selector and select Analytics.
2. Click Language Model > Categories.

3. Click more options in the row of the group you want to manage. Select the action that best suits your needs.

   Learn more about actions in this step

   Field	Details
   Rename	Enables you to rename the folder.
   Move	Enables you to move a category Data groups that Interaction Analytics uses for parsed call transcript data. They make it easier to find trends. to another folder or location. To rearrange the categories, you can also click and hold while dragging the category to where you want it.
   Duplicate	Enables you to copy a category. This is helpful if you want to create a category similar to an existing one with slight modifications.
   Delete	Deletes the category.
   New Category	Creates a new category.
   New Group	Creates a new category folder.
   Import Out of Box Category	Enables you to import an existing category or category folder with preconfigured rules.

Manage Category Rules

To create specific criteria an interaction must meet to fall into the category Data groups that Interaction Analytics uses for parsed call transcript data. They make it easier to find trends., you can modify the category rules Criteria used to determine whether a transcript is placed in a certain category. Rules contain sets of keywords or phrases a transcript must include or exclude.. The rules can require a certain sentiment Overall mood or result of the interaction as determined by analysis of words, phrases, and context of the transcript. or its absence, or certain keywords or phrases Combination of words that have special significance when used together in a specified order, such as "want to cancel". or their absence. You can specify whether your rule is looking for beginning, overall, or end sentiment. This allows you to determine whether a contact The person interacting with an agent, IVR, or bot in your contact center. was positive or negative at any time during the interaction The full conversation with an agent through a channel. For example, an interaction can be a voice call, email, chat, or social media conversation.. You can also create rules that use other metrics and filter criteria. Rules can be large and detailed or they can be broad and simple.

You can include or exclude interactions that meet some or all of the criteria you enter. You can also specify at which point in the interaction certain criteria occurred.

If you create multiple category rules for a category, the interaction must match only ONE of the category rules. It does not need to match all category rules to be included.

1. Click the app selector and select Analytics.

2. Click Language Model > Categories.

3. Click the category you want to modify.

4. In the Timing field, specify the time period in which the keywords, phrases, or entities have to occur in the category rule.

   Learn more about fields in this step

   Condition	Details
   Any time during interaction	Criteria can occur at any point during an interaction.
   When all phrases are found within	All phrases must be found in an interaction within a certain amount of seconds between them. For example, if you include three phrases and set the value for 30 seconds, all phrases must be mentioned within 30 seconds of each other.
   Between interaction times	You can specify a specific time period during an interaction in seconds. Enter when you want the category to start and stop looking for the criteria. For example, you could set the category to look for criteria from 20 seconds into the interaction to 80 seconds into the interaction.
   At the beginning of the interaction	The first 400 words of the interaction or the first 30%, whichever occurs first.
   At the end of the interaction	The last 30% of the interaction.

5. In the Participant field, specify whether you want the rule to apply to the Agent side of the interaction, the Client (contact) side of the interaction, or to both the Agent and Client side of the interaction.

6. Select the conditions you would like the criteria to meet for each field. You can include or exclude interactions that meet some or all of the criteria you enter. All keywords and phrases entered in the same field have an AND relationship, and the fields have an AND relationship between them. This means that all the specified conditions of a single rule set must exist for an interaction to fall into the category, except those in the At least one of these field.

   Learn more about fields in this step

   Condition	Details
   All of These	An interaction must include all keywords, phrases, or tags entered in this field to be a match. All the criteria in the field assume an AND relationship.
   At Least One of These	An interaction may be a match if it includes one of the keywords, phrases, or tags entered in this field. You can create multiple instances of this field by pressing the + button. All criteria within a field assume an Or relationship to each other. However, each instance of the field assumes an AND relationship with the others. For example, if one At Least One of These field contains keywords Longbourn and Bennet, interactions containing one or both of those words are included. If another At Least One of These field contains keywords Netherfield and Bingley, interactions that mention either of those words are also included, even if they don't mention the keywords Longbourn or Bennet from the other criteria field. Interactions that contain any combination of these keywords from both criteria fields would be included.
   None of These	Interactions must not contain any one of the keywords, phrases, or tags entered in this field to be a match.

7. To include or exclude interactions by Phrase, Keyword, or Sentiment , click the corresponding tag or begin typing a keyword, phrase, or entity Keyword or phrase defined in your company profile in Interaction Analytics. Related to an entity type. Can include variants..

   Learn more about fields in this step

   Field	Details
   Phrase	Pulls in data based on whether a specific phrase Combination of words that have special significance when used together in a specified order, such as "want to cancel". was mentioned in interactions. Type a phrase and click Enter or select the Phrase tag. Phrases must be enclosed in quotation marks, for example, "Rosings Park". Click the button to the right of the phrase to search for an Exact Match or to Match with Variations. = Exact Match is the default. If you choose Match with Variations, the criteria will also include variations similar to the phrase you entered. For example, if you enter "want to cancel", it could return interactions that include wanted to cancel, wanting to cancel, and so on. Click the Proximity drop-down to specify how close the words in the phrase must be to each other for an interaction to be considered a phrase match. The default is 0, meaning the words must be together and in the right order. Configuring Proximity means the words in the phrase don't have to be consecutive and that they can have up to eight words between them. For example, the phrase speak manager with a proximity of 2 would capture the phrases "speak to a manager", "speaker with your manager", and so on.
   Keyword	Pulls in data based on whether a specific keyword was mentioned in interactions. Type a keyword and click Enter or select the Keyword tag. Click the button to the right of the phrase to search for an Exact Match or to Match with Variations. = Exact Match is the default. If you choose Match with Variations, the criteria will also include variations similar to the phrase you entered. For example if you enter activate, it could return interactions that include activate, activates, activating, activated, and so on.
   Sentiment	Pulls in data based on the sentiment Overall mood or result of the interaction as determined by analysis of words, phrases, and context of the transcript. assigned to interactions. Select the Sentiment you want to focus on: Positive, Negative, Neutral, or Mixed. Click the double arrows to select when this sentiment must have occurred during the interaction. Overall sentiment is the default. Overall : The sentiment for the whole interaction. Beginning Sentiment: This is determined by the first 400 words, or first 30%, of the interaction, whichever occurs first. It can help you identify common reasons or problems clients call about. Mixed sentiment is not available for beginning sentiment. End Sentiment: End sentiment is determined by the last 30% of an interaction. Mixed sentiment is not available for end sentiment.

8. To include or exclude interactions by metrics like team name or ACD skill name, click the Metric tag and select the metric type from the list.

   Learn more about fields in this step

   Field

   Details

   Metric

   Pulls in data based on whether it meets a certain metric value or threshold. See the Metrics entry in Key Terms and Metrics for more information about each available metric type. Depending on the metric you select, you need to complete one of these additional configuration steps: If you select a boolean field, select true from the drop-down to show only interactions where the field applies. Select false to show only interactions where the field does not apply. If you select a string field like a name or ID, enter or select the value in the text field. This will only include interactions where the specified condition exists. If you select an integer field, like seconds or counts, select an operator from the drop-down and enter an integer value. Only interactions that meet the specified threshold appear in the dataset. Equals: Data appears in the dataset if the value for the field is equal to the specified value. Greater Than: Data appears in the dataset if the value for the field is higher than the number specified. Less Than: Data appears in the dataset if the value for the field is lower than the number specified

   If you want to filter for a range of values, you can add multiple of the same kind of metric filters with different operators and values. For example, if you want to include interactions that are between two and four minutes, you would select the interactionDurationSeconds metric. Configure it to > 120. Next, add a second interactionDurationSeconds metric and configure it for < 240. Then, change the condition to All of these for the category rule.

9. To include or exclude interactions by Frustration or Resolution, select the corresponding tag.

   Learn more about fields in this step

   Field	Details
   Frustration	Pulls in data from interaction where the contact became frustrated Looks for cues to identify customer frustration. The cues include words and phrases like, "I'm very angry". Frustration isn't the same as negative sentiment. Frustration cues show that a customer is upset rather than just discussing something negative..
   Resolution	Pulls in data based on the resolution status at the end of a contact. Select Resolved to view only resolved interaction data. Select Unresolved to view only unresolved interaction data.

10. To add a new rule set, click Add Rule. Rule sets can have one of three types of relationships:
    • OR: Interactions don't have to meet all the rules to be included in the category. So, interactions that satisfy any one of the rule sets configured in the category rule fall into the category. This is the default option, but you can change the configuration option by clicking OR, and selecting another option from the drop-down.
    • AND: An interaction must meet all rule criteria to be included in the category.
    • NOT: The interaction must meet the rule criteria in the previous set. It cannot meet the criteria in the following rule set.

    You can group rules and rule sets together to create more complex rules.

    Learn more about grouping rules

    You can create two or more rule sets and group them together using AND, NOT, and OR within each rule set. You can then have additional rules or rule sets and define how each one interacts with the other. This allows you to create more complex rules to determine how you want interactions to match. For example:

    Without Grouping

    Rule1 AND Rule2 OR Rule3​

    AND is evaluated first, then the OR​

    Results:​

    Rule1 AND Rule2 = True (match)​

    OR Rule3 = True (match)

    With Grouping

    Rule1 AND (Rule2 Or Rule3)​

    Forces Rule2 and Rule3 to be evaluated first, and then is paired with Rule1​

    Results:​

    If Rule2 OR Rule3 is a match, it will now look at Rule1​

    If Rule1 is a match, then the interaction matches the category​

    If neither Rule2 OR Rule3 is a match, no match​​

11. When you are finished adding criteria, click Save.

Resample Data and Recategorize

You can pull in a sample set of interactions The full conversation with an agent through a channel. For example, an interaction can be a voice call, email, chat, or social media conversation. to see how the current categories Data groups that Interaction Analytics uses for parsed call transcript data. They make it easier to find trends. are working. This helps you see how your changes affect the sample data before you apply those changes to a category set Category groupings that help you use the data for a specific purpose. Out-of-the-box category templates include Intent to Buy and Risk Aversion.. You can then recategorize based on the results of the sampled data.

1. Click the app selector and select Analytics.
2. Click Language Model > Categories.

3. Either create a custom category set or modify an existing one.

4. Click Resample . It loads a fresh sample set of interactions and categorizes them. It can take some time.

   Do not click Resample more than once. It can look like nothing is happening for a moment, but a progress bar appears to show you the status. Clicking the button multiple times causes the system to queue up the actions and load and reload contacts for a long time.

5. Perform any of the actions documented on this page to modify the category tree to meet your needs.
6. Click Recategorize . This causes your existing data samples to reflect the changes you made to the category set.