HELP FILE

What are voices?

Voices are a visual representation of the questions your customers are asking on Bold360 ai self-service engagements. With Voices, you gain valuable insight into customer sentiment by tracking question trends, answer feedback and problem resolution.

Voices are a visual representation of the questions your customers are asking on Bold360 ai self-service engagements. With Voices, you gain valuable insight into customer sentiment by tracking question trends, answer feedback and problem resolution.

How are Voices collected?

Every time your customers asks a question or looks up information on Bold360 ai's self-service touch points, a customer interaction or inquiry is collected. Bold360 ai's algorithm clusters these inquiries into phrases and the phrases into voice clusters. Bold360 ai keeps track of all of these voice and displays them as tiles on your dashboard. The bigger the tile, the more popular that topic is in your customer' voices.

Users often ask the same questions in many different ways. Bold360 ai groups similar questions together to present a clear picture the different topics your customers are asking about. Click on a voice to see the different phrasings your customers use to ask that question.

Clicking a voice phrasing from within the search optimizer shows the phrasing details, including the current answer provided by Bold360 ai as well as any feedback information.

A voice is mapped to an article via the article title and the phrases that underpin an article.

How do Voices reflect customer sentiment?

Your customers are encouraged to leave feedback during their engagement with Bold360 ai self-service channels. If they like the answer, the voice 'like' counts increases. If they dislike something they'll have to chose between several reasons. Voices with negative feedback are counted as unresolved.

in Voices, Bold360 ai shows customer feedback using tile colors: blue tiles represent neutral sentiment and correspond to resolved engagements, red tiles represent negative sentiment and correspond to unresolved engagements, green tiles represent extremely positive feedback. For more depth on the algorithm that defines the sentiment level and how to configure it click here.

How are voices classified?

  • A voice that is processed and associated with an article is neutral and therefore blue.
  • A voice that is processed and remains unanswered is classified as a red voice, calling for attention.
  • A voice can also be colored red if it receives an amount of negative feedback that is above the account's health threshold and the minimum amount of interactions with feedback.
  • A voice will be classified as green if the amount of positive feedback is above the health threshold and the minimum amount of interactions with feedback.
  • A voice that has both significant negative and positive feedback will be colored red, as it signals mixed opinions and therefore needs to be explored.

The account health settings can be found on the Admin Center > KB Settings > Account health tab. The default values are based on Bold360 ai's analytics across all accounts. It may be useful to check your customers' feedback propensity and modify the values accordingly.

The following example illustrates a voice classification life cycle based on the account health parameters defined above.

  • The default health threshold is defined as 0.5 (amount of feedback divided by the amount of interactions)- meaning: to get a color change you need 50% negative (or positive) feedback.
  • The minimum amount of feedback that is needed in order for the classification to be considered: 2 interactions.
Week Interactions Feedback Voice Color Classification
1 100 1 negative Blue (negative FB above threshold but below minimum amount of interactions with FB)
2 150 2 negative Red
3 300 2 negative2 positive Red (the voice is classified as red if the amount of negative FB suggests an issue, regardless of the amount of positive FB)
4 500 2 negative2 positive Blue (neither negative nor positive FB reaches the threshold)
5 600 2 negative4 positive Green (negative FB did not reach the threshold and positive FB above threshold)

It is important to note that the color classification is cumulative across time and is not related to the time period selected for the dashboard.