AI Analytics Articles

How to monitor traffic on messaging channels

You can monitor the performance of your WhatsApp and Microsoft Teams messaging channels in the AI Console as well as the Agent Workspace.

Monitoring capabilities in the AI Console

You have the following options to run reports and monitor messaging conversations:

  • Engagement Dashboard: provides a view on the total number of engagements, the number of engagements escalated to agents and so on. See What do I see on the Engagements Dashboard? for more information.
  • Dashboard: You can look at voices for the knowledge base you have for messaging, or filter the dashboard using the context you use for messaging to see how your messaging channel is doing. See What do I see on the Dashboard? to learn more.

Monitoring capabilities in the Agent Workspace

The following options are available to run reports and monitor messaging conversations:

  • You can view active bot, queued, assigned to agent as well as closed messaging conversations in the Monitor View of the Agent Workspace.
  • You can run reports on messaging conversation data with Data Extraction API methods. For more information, see the Developer Center.
  • You can run status and summary reports on messaging conversations. For more information see Messaging Apps Reports.

What are the Support Center Analytics?

The Support Center is also trackable with the Bold360 AI Analytics functionality.

From within your Bold360 AI account, go to Analytics > Custom Charts > Support Center.

The following parameters are available for analytics tracking:

  • Visits
  • Page Loads
  • Organic Visits (visits generated via search engines rather than paid advertisements)
  • Total resolved by Support Center

Knowledge Base Analytics: Most Escalated, Most Popular, Most Liked/Disliked

To view Knowledge Base analytics, go to Analytics > Knowledge Base.

Most Popular
The most commonly asked questions.
Most Escalated
The questions that most often lead to escalations.
Most Liked
Questions that received the most "like" responses.
Most Disliked
Questions that received the most "dislike" responses.
Tip:
  • Click any question to see a graphic visualization of a question's trend over time.
  • Use the search box (by keyword) to filter by any specific question. The filter applies to all four charts.
  • When applicable, Context is shown in parentheses.

What is the Knowledge Base Usage Chart

The Knowledge Base Usage chart shows you how many questions were added or deleted in the knowledge base over a period of time.

To view the chart, in Bold360 AI, go to Analytics > Knowledge Base.

The Knowledge Base Usage Chart provides you with a brief overview of how customer support has been operating over a specified period of time. This data enables you to ascertain how your customer team is functioning.

The legend for the Knowledge Base Usage Chart is as follows:

  • The green line indicates the number of answers that were added to the Knowledge Base.
  • The blue line indicates the number of answers deleted.

See also: Knowledge Base size Chart

What is the Knowledge Base Size Chart?

To view the Knowledge Base Size Chart, in Bold360 AI, go to Analytics > Knowledge Base.

The Knowledge Base Size Chart shows you the volume of questions (including phrasings) and answers in the knowledge base. It also displays the volume of tickets in the inbox and the volume of items in the publishing center under the publish answers tab on a time scale.

All these metrics give you the progress of the system's learning process and knowledge base growth.

Notes:

  • The yellow line shows the number of questions and phrasings
  • The red line shown the number of answers.
  • The blue line shows the number of items in the publishing center under the publish answers tab
  • The green line shows the number of tickets in the inbox

See also: What is the Knowledge Base Usage Chart

How can I report that I found personal data?

Digital DX AI has a built-in mechanism to replace personal data with asterisks in user searches, but sometimes you may still find personal data in Search Optimizer, on the Dashboard or in Reporting.

In the AI Console, users with view permission for Analytics have the option to mark user queries that contain personal data so that they can be removed. You have modify permission for Analytics? Follow these steps to obscure and remove the data.

  1. Go to Analytics > Reports.
  2. Run a report that contains the Query field.
  3. Find the query that contains sensitive data and choose (Personal Informaition in Query).
The icon turns red and the value of the Holds PII column changes to True for the query.

How do I add a tag data to conversion tracking?

Note: Conversion tracking has been integrated into Bold360 Acquire. For more information, contact your Customer Success Manager.

The following API allows you to identify the user in conversion reports in order to track the conversion on your website. It also measures the impact that Bold360 AI usage has on conversion (showing the conversion gain for users that use the Bold360 AI widget).

The conversion tracking tool includes a Conversion Log that shows the time and IP address of each converted user that used the Bold360 AI widget.

You can add tagging for each user that will be shown in the conversion log for each conversion.

This can be done by adding the conversionTag value.

Example:

Replace CONVERSION_TAG with a unique page name.
_nRepData["conversionTag"] = "CONVERSION_TAG";
Example:

To identify a user who purchased a camera, add this line:
_nRepData["conversionTag"] = "Camera";

How can I add conversion tracking?

Note: Conversion tracking has been integrated into Bold360 Acquire. For more information, contact your Customer Success Manager.

Here's how to add a new journey (also known as conversion tracking script):

  1. Go to Analytics > Tracking.
  2. Click New Journey Tracking at the bottom of the page.

    Result: The New Journey page is displayed.

  3. On the New Journey page, enter a name and click Next.

    Result: The [New Journey Name] page is displayed.

  4. Set up your journey script:
    1. Track visitors' URL - Choose the domain to which you want to add the journey script and collect journey analytics.
      Tip: To add the journey script to all of your domains, select All Domains.
    2. Track my sales revenue - Include sales revenue in your journey script. You must choose a currency.
      Important: Revenue statistics/analytics will be unavailable unless you enable this option.
    3. Split test control group - Specify the percentage you want to measure within the control group - that is, the number of visitors who will not see a Bold360 AI widget.
  5. Click Next.

    Result: The conversion tracking code snippet is displayed. Here's a sample code snippet:

    <script type='text/javascript' src='https://my.Nanorep.com/~YOUR_ACCOUNT_NAME/common/API/trackConversion.js? account=YOUR_ACCOUNT_NAME&id=JOURNEY_ID&sid=39316202&revenue=TOTAL&orderId=ORDER_ID'> </script>
  6. The code contains placeholders. You must replace these with real values to correctly track your sales:
    1. Replace TOTAL with your current sale value in your currency. For example, revenue=50.00
    2. Replace ORDER_ID with your order ID. The order ID parameter is a dynamic parameter, based on the order ID on your site.

Conversion tracking troubleshooting

Note: Conversion tracking has been integrated into Bold360 Acquire. For more information, contact your Customer Success Manager.

If you cannot enable the conversion tracking script generated by Bold360 AI in your website, do the following.

  1. Go to Analytics > Tracking.
  2. Hover over your conversion tracking entry and click the Navicon menu (three lines).
  3. Select Configure.

Possible errors for the split test:

No traffic from conversion, No conversions yet.
Solution:
  • Make sure your tracking code is correctly installed on your website conversion page.
  • Make sure your conversion code is posted in the post purchase page
  • Make sure you replace placeholders with valid numbers

Example:

<script type='text/javascript' 
src='https://server3.Nanorep.com/~trueshopping/common/API/trackConversion.js?
account=trueshopping&id=36073339&sid=36073344&revenue=40.00&orderId=12345678' 
</script>
  • revenue = Current sale revenue in your currency
  • orderid = Your order ID.
Split test set to 0%, Split test isn't running.

Solution: Set the split test percentage to a higher value. You can increase the split test percentage by selecting the required value from the Split test control group drop-down list in your journey's settings page.

Why don't I see the revenue tracking?

Make sure that Track sales revenue is enabled in your conversion tracking.

  1. In the AI Console, go to Analytics > Tracking.
  2. Select a journey that you use for tracking.
  3. Select an option from the Track sales revenue drop-down list.

To learn more about the Manage Conversions page, click here.

In case of any issues check our Conversion tracking troubleshooting guide.

Using the Analytics API

The Analytics API enables you to access and collect data from the Bold360 AI Analytics section.

You can get a more comprehensive picture of your customer support processes and self-service success. You can also build custom reports that show your customer support impact.
  1. To use the Analytics API, create an API key in your Bold360 AI account.

    For information on how to create a new API key, click here.

  2. To retrieve analytics data with the API, use the get method.
    • For a JSONP response, use: https://my.nanorep.com/api/analytics/v1/get
    • To retrieve a CSV file, use: https://my.nanorep.com/api/analytics/v1/get.csv
    Table 1. Input parameters
      Usage Example
    cb

    The JSONP callback method.

    Note: JSONP only.
    cb=callbackMethod
    account

    The account name.

    account=myAccount
    apiKey

    An API key with analytics read privileges for the required knowledge-bases.

    apiKey=xxx-yyy-zzz
    kb

    One or more knowledge-base names to obtain data for (comma-separated).

    kb=English,Spanish
    analysis

    Value analysis type. Possible values are:

    • hourly
    • daily
    • weekly
    • monthly
    analysis=weekly
    days
    Note: You must provide either a day or month value when selecting a date range.

    Date range in days to obtain data for.

    days=14
    months
    Note: You must provide either a day or month value when selecting a date range.

    Date range in months to obtain data for.

    months=6
    timezone

    Timezone in hours to obtain correct dates.

    timezone=5
    values

    List of values to include in output (comma-separated).

    List of possible values is provided in the table below.

    values=engagements,resolved
    labelid (Optional) Get data on specific label. labelId=1CE001C
    Table 2. Possible values for values parameter
    Value name Description

    Engagements

    answeredSessions Total answers provided, including self-service, rechanneling, and tickets.
    interactions Total interactions via Bold360 AI interfaces (widgets, Support Center, API, and so on).
    answered Interactions where search has returned an answer.
    unanaswered Interactions where search did not return an answer.
    resolved Engagements resolved by Bold360 AI (visitor received an answer, and did not escalate or provide negative feedback).
    unresolved Engagements not resolved by Bold360 AI (either no search result has been found or the visitor's question was escalated to an agent, or the visitor provided negative feedback).
    resolvedBySearch Engagements resolved using the Search interface.
    resolvedBySearchAC Engagements resolved using the Search interface by selecting an auto-completion suggestion (subset of resolvedBySearch).
    resolvedByFAQ Engagements resolved using an FAQ interface (in Dynamic FAQ widget, or FAQs in embedded or floating widgets).
    resolvedBySupportCenter Engagements resolved using a Support Center article.
    positiveFeedback Number of positive feedback from all touchpoints.

With Bold360, you never have to guess how you’re doing. Key KPIs and actionable insights are right in front of your eyes with overviews and snapshots. Or drill down into key KPIs to discover actionable insights. Detailed report results pinpoint when customers are experiencing friction, so you know where to take action. With Bold360, you have control over every rule and setting, so you can make any changes that you see fit to improve your business metrics.

Can I export analytics data from the SLA page?

Yes, you can use the Export option near Label Statistics page in console = > Analytics = > SLA menu

How can I measure my customers' satisfaction from articles?

To get estimations on the quality of your articles or a single article, you can create a report. The report will show the number of likes compared to total likes and dislikes of the article.

To create the report, do the following:

  1. In the Bold360 AI platform, go to Analytics > Reports and choose the following values:
    1. Under Fields category, select: Knowledge base, Article ID, Date & Time, Article Title, #Positive FB, Negative FB, #Interactions.
    2. Under Group by category, select: Knowledge base, Article ID.
  2. Choose the time period you would like to view.
  3. Click Run.

Results are sorted by Knowledge base.

To view more details, click on the arrow next to the Article ID.


How to generate a report on interface sources?

A report on interface sources tell you where your customers receive answers from.

To run an interface source report, do the following:

  1. In the AI Console, go to Analytics > Reports.
  2. On the Fields tab, select Interface as Source.
  3. Select any other fields that you want to include in the report.
  4. On the Group by tab, select Interface so that your results are grouped by interface sources in the report.
  5. Run the report.

The interface report groups interfaces into the following categories:

  • FAQ: interactions in the FAQ section of a support center or an on-journey widget
  • Search: interactions, where the customer used search without auto-complete
  • Auto-Complete: interactions where the customer used search, and selected one of the proposed answers instead of pressing Enter
  • Support center: interactions in the navigation area of the support center. This excludes interactions in the search widget of the support center.
  • Linked Article: interactions where the customer clicked on a linked to read another article
  • API: interactions that were served directly via an API rather than a Bold360 AI touchpoint
  • Onload: interactions where a welcome message was displayed first
  • Bot: interactions that were served on the bot through search or auto-complete
  • Guidance: not in use

How can I generate my own reports in Bold360 AI?

Bold360 AI allows you to easily create your own custom reports to retrieve data regarding the traffic in Bold360 AI. You can select a wide range of fields to show in the report, grouped them, filter them by a certain criteria and more. To generate your own reports in Bold360 AI, do the following:

  1. In the AI Console, go to Analytics > Reports.
  2. On Fields tab, select the fields that you want to include in your report. Move your mouse over a field to see more information.
    Note: You can select up to 15 fields for one report.
  3. On the Grouped by tab, select fields that will group your report data. For example if URL is chosen, the report will show you information grouped by URL.
    Note: You can select up to two fields to group your data.
  4. On the Filter by tab, select specific values to filter the attributes in the report.

    For example, if the report retrieves data regarding "Channeled type" in the filter tab, you can select to filter the "Channeled type" to show only a specific type of channel, such as "Chat".

  5. On the Sort by tab, select the sorting order of the report fields.
  6. Once all the attributes in the report are set, select the date range for the report and click Run to generate the report.

The report is displayed on the same page, and you can download it as a CSV file. The downloaded file contains a detailed report. If you want the summary view as presented on the screen, select Export view to CSV.

Note: Data in the reports is generated in UTC time zone.

To save the report for future use, click Save and define the name of the report.

How to run reports on article labels

In the definition of Reports, on the Filter by tab, labels are presented under their respective knowledge base. This simplifies label selection when labels are named similarly across knowledge bases.

How to understand origin of queries

When selecting report fields, you can include the Domain and Full URL fields to report on the origin of particular queries.

What do I see on the Engagements Dashboard?

The Engagements Dashboard shows the total number of engagements, engagement-related metrics, and offers a detailed view on individual engagements.

By reviewing experiences that share an outcome, for example, channeled engagements, you may gain insights into how to further drive optimization.

With this dashboard you can gain insights into the following:

  • Total number of engagements
  • Engagement self-service score
  • Deflection rate
  • Engagement trend
  • Other details about engagements such as the amount of channeled engagements, engagements with positive or negative feedback and so on
  • Details about the user?s journey such as the date and time of the engagement, the event type, and the user query

You can do the following with the data on this dashboard:

  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Engagement Summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Engagements summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Engagements summary table
  • Select a number or a percentage in a cell in the Engagements summary table to see the details of the user?s journey for each engagement

What is an engagement?

An engagement is a collection of end user events with the bot, with a live agent or both where the longest interruption between two events is no longer than 30 minutes.

An engagement is a user's visit in which a Bold360 capability was used at least once. An engagement may include a variety of events incorporating both Bold360 AI and Bold360 Agent events. Such events can be typing a question, selecting an article, messaging with a live agent, or filling in a chat feedback form. An engagement ends after 30 minutes of idle time.

The engagement calculator service is continuously monitoring incoming raw events and determines if an event is part of an active engagement or belongs to a new one.

What is a proactive engagement?

Proactive engagements are triggered by proactive strategies or campaigns. They are not initiated by the visitor or the agent.

What is the engagement self-service score and how is it calculated?

Engagement self-service score indicates whether the user was able to achieve their objective through self-service. It is calculated by dividing the number of engagements without negative signals with the total number of engagements.

An engagement is self-served if it meets the following criteria:

  • It doesn't include a channeling event
  • It doesn't include a negative feedback event
  • The bot was able to answer at least one user question

The engagement self-service score is expected to be lower than the interaction self-service score. Both engagements and interactions are a certain type of event grouping, with different levels of granularity. In practice, each engagement includes one or more interactions. Therefore, a single negative signal for an interaction has higher impact on engagement self-service score than on interaction self-service score.

For example, if an engagement with five interactions receives a single negative signal, we get 0% engagement self-service score and 80% interaction self-service score.

What is the deflection rate and how is it calculated?

Deflection rate shows the extent to which automated and self-help resources satisfy service demand that would otherwise be handled by support staff. It is calculated by dividing engagements without channeling events with the total number of engagements and is shown as a percentage.

What data does the Engagement Summary table contain and how are they calculated?

The Engagement Summary table shows the following pieces of information about engagements:

  • Channeled: percentage of engagements that contain at least one channeling event like chat, ticket, or phone.
  • Positive feedback: percentage of engagements that contain at least one positive feedback event.
  • Negative feedback: percentage of engagements that contain at least one negative feedback event.
  • Channeled to agent: percentage of engagements that are channeled to an agent.
  • Agent only: percentage of engagements that include agent live chat only.

What time interval is used on this dashboard?

Engagement data for the current month and the previous three months is available on this dashboard. The exact time interval is displayed under the Engagement Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated one a day. The last update time is displayed at the top, above the Engagement trend widget.

What do I see on the Interactions Dashboard?

The Interactions Dashboard shows the total number of interactions, interaction-related metrics, and offers a detailed view on individual interactions.

Note: The Interactions Dashboard is in beta. Contact your Customer Success Manager for more information.

With this dashboard you can gain insights into the following:

  • Total number of interactions
  • Interaction self-service score
  • Interactions trend
  • Helpful rate
  • Other details about interactions such as the amount of channeled interactions, interactions with positive or negative feedback and so on
  • Details about the user?s journey such as the date and time of the interaction, the event type, and the user query

You can do the following with the data on this dashboard:

  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Interactions summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Interactions summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Interactions summary table
  • Select a number or a percentage in a cell in the Interactions summary table to see the details of the user?s journey for each interaction

What is an interaction?

An interaction represents a user?s attempt to receive information or issue a request as it relates to a specific topic and includes all follow-up actions around the information retrieved. There are two types of interactions: AI-centric interactions and human-assisted interactions.

In an AI-centric interaction, a search event represents the start of an interaction and any follow-up events such as channeling, feedback and conversation events are included up to a period of 24 hours. An interaction conveys the treatment of a specific user need or intent, therefore it will include any piece of follow-up information that is needed to respond to the user's request.

A human-assisted or agent-facing interaction is comprised of one or more end user utterances and the agent?s response or responses.

Important: A follow-up event generated outside the 24-hour period will generate a new engagement, and will not be included in the previous interaction. It will count as an individual event. The first agent message without a prior visitor message, such as a welcome message, doesn't create a new interaction.

What is the interaction self-service score and how is it calculated?

The interaction self-service score indicates whether the user was able to find a response to a specific request. It is calculated by dividing the number of interactions without negative signals with the total number of interactions.

An interaction is self-served if it meets the following criteria:

  • It doesn't include a channeling event
  • It doesn't include negative feedback
  • The bot was able to answer at least one user question

What is the helpful rate and how is it calculated?

Helpful rate shows the percentage of positive feedback on interactions. It is calculated by dividing the number of positive feedback with the total number of feedback.

What is the feedback rate and how is it calculated?

Feedback rate shows the percentage of feedback on interactions. It is calculated by dividing the total number of feedback with the total number of interactions.

What data does the Interactions Summary table contain and how are they calculated?

The Interactions Summary table shows the following pieces of information about engagements:

  • Channeled: percentage of interactions that contain at least one channeling event like chat, ticket, or phone.
  • Unanswered: percentage of interactions that are not answered.
  • Positive feedback: percentage of interactions that contain at least one positive feedback event.
  • Negative feedback: percentage of interaction that contain at least one negative feedback event.
  • Channeled to Bold Agent: percentage of interactions that are channeled to an agent.
  • Interactions with Bold Agent: percentage of interactions that include agent live chat only.

What time interval is used on this dashboard?

Interaction data for the current month and the previous three months is available on this dashboard. The exact time interval is displayed under the Interaction Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated one a day. The last update time is displayed at the top, above the Interaction trend widget.

What do I see on the Channeling Dashboard?

The Channeling Dashboard shows the total number of interactions, interaction-related metrics, and offers a detailed view on individual interactions.

Note: The Channeling Dashboard is in beta. Contact your Customer Success Manager for more information.

The Channeling Dashboard allows users to monitor customer?s channel preferences, optimize channel distribution, and continuously reduce the channeling rate of low-value transactions while providing better self-service support for these interactions.

With this dashboard you can gain insights into the following:

  • Channeling trend
  • Channeling distribution
  • Top 10 channeled articles
  • Channel value
  • Details on each channeled interaction such as channel type, value and name, the number of negative feedback and whether it was unanswered

You can do the following with the data on this dashboard:

  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Channeling summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Channeling summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Channeling summary table

What does the channeling trend chart show?

The Channeling Trend chart shows the number of channeled interactions for each day per channel type (chat, email, ticketing or phone).

What does the channeling distribution chart show?

The Channeling Distribution chart shows the distribution of channel types (chat, email, ticketing or phone) as compared to the total number of engagements for the selected period.

What does the channel value show?

The Channel Value chart shows the distribution of channel sentiments for the selected period.

Channel Sentiment is the negative, neutral or positive score assigned to a channel in the channeling policy. See What is intentional channeling? to learn more.

What data does the Channeling Summary table contain and how are they calculated?

The Channeling Summary table contains the number of channeled interactions grouped by knowledge base, channel type and name.

The Channeling Summary table shows the following pieces of information about articles:

  • Channel value: shows the channel sentiment as defined in the channeling policy.
  • Interactions: number of interactions for the channel
  • Has negative feedback: number of interactions with negative feedback.
  • Unanswered: number of interactions that are not answered.

What time interval is used on this dashboard?

Interaction data for the current month and the previous three months is available on this dashboard. The exact time interval is displayed under the Channeling Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated one a day. The last update time is displayed at the top, above the Channeling trend widget.

What do I see on the Engagements History Dashboard?

The Engagements History Dashboard shows the total number of engagements, engagement-related metrics, and offers a detailed view on individual engagements for the last 12 months.

Note: The Engagements History Dashboard is in beta. Contact your Customer Success Manager for more information.

By reviewing experiences that share an outcome, for example, channeled engagements, you may gain insights into how to further drive optimization.

With this dashboard you can gain insights into the following:

  • Total number of engagements
  • Engagement self-service score
  • Deflection rate
  • Engagement trend
  • Engagement type trend
  • Customer effort score
  • Other details about engagements such as the amount of channeled engagements, engagements with negative feedback and more

You can do the following with the data on this dashboard:

  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Engagements Summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Engagements Summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Engagements Summary table
  • Select a number or a percentage in a cell in the Engagements Summary table to see the details of the user's journey for each engagement

What is an engagement?

An engagement is a collection of end user events with the bot, with a live agent or both where the longest interruption between two events is no longer than 30 minutes.

An engagement is a user's visit in which a Digital DX capability was used at least once. An engagement may include a variety of events incorporating both AI and Agent events. Such events can be typing a question, selecting an article, messaging with a live agent, or filling in a chat feedback form. An engagement ends after 30 minutes of idle time.

The engagement calculator service is continuously monitoring incoming raw events and determines if an event is part of an active engagement or belongs to a new one.

What do the Engagement Trend and Engagement Type Trend charts show?

The Engagement Trend chart shows the total number of engagements per month as well as the deflection rate and the self-service score.

The Engagement Type Trend chart shows the number of AI-only engagements and the engagements that include both AI and agent live chat events.

What is the engagement self-service score and how is it calculated?

Engagement self-service score indicates whether the user was able to achieve their objective through self-service. It is calculated by dividing the number of engagements without negative signals with the total number of engagements.

An engagement is self-served if it meets the following criteria:

  • It doesn't include a channeling event
  • It doesn't include a negative feedback event
  • The bot was able to answer at least one user question

The engagement self-service score is expected to be lower than the interaction self-service score. Both engagements and interactions are a certain type of event grouping, with different levels of granularity. In practice, each engagement includes one or more interactions. Therefore, a single negative signal for an interaction has higher impact on engagement self-service score than on interaction self-service score.

For example, if an engagement with five interactions receives a single negative signal, we get 0% engagement self-service score and 80% interaction self-service score.

What is the deflection rate and how is it calculated?

Deflection rate shows the extent to which automated and self-help resources satisfy service demand that would otherwise be handled by support staff. It is calculated by dividing engagements without channeling events with the total number of engagements and is shown as a percentage.

What is the customer effort score and how is it calculated?

Customer effort score shows the average number of interactions it takes to conclude an engagement. It is calculated by dividing the number of interactions with the number of engagements.

What data does the Engagements Summary table contain and how are they calculated?

The Engagements Summary table shows the following pieces of information about engagements:

  • Engagements: total number of engagements for the knowledge base
  • Bot engagements: total number of AI-only engagements for the knowledge base
  • Channeled: percentage of engagements that contain at least one channeling event like chat, ticket, or phone.
  • Channeled to agent: percentage of engagements that are channeled to an agent.
  • Agent only: percentage of engagements that include agent live chat only.
  • Self-service score: the percentage of engagements without negative signals divided by the total amount of engagements.
  • Customer effort score: number of interactions divided by the number of engagements.

What time interval is used on this dashboard?

Engagement data for the last 12 months is available on this dashboard. The exact time interval is displayed under the Engagements Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated once a day. The last update time is displayed at the top, above the Engagement trend widget.

What do I see on the Interactions History Dashboard?

The Interactions History Dashboard shows the total number of interactions, interaction-related metrics, and offers a detailed view on individual interactions for the last 12 months.

Note: The Interactions History Dashboard is in beta. Contact your Customer Success Manager for more information.

With this dashboard you can gain insights into the following:

  • Total number of interactions
  • Interaction self-service score
  • Interactions trend
  • Interactions type trend
  • Other details about interactions such as the amount of channeled interactions, interactions with negative feedback and more

You can do the following with the data on this dashboard:

  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Interactions summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Interactions Summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Interactions Summary table
  • Select a number or a percentage in a cell in the Interactions summary table to see the details of the user?s journey for each interaction

What is an interaction?

An interaction represents a user?s attempt to receive information or issue a request as it relates to a specific topic and includes all follow-up actions around the information retrieved. There are two types of interactions: AI-centric interactions and human-assisted interactions.

In an AI-centric interaction, a search event represents the start of an interaction and any follow-up events such as channeling, feedback and conversation events are included up to a period of 24 hours. An interaction conveys the treatment of a specific user need or intent, therefore it will include any piece of follow-up information that is needed to respond to the user's request.

A human-assisted or agent-facing interaction is comprised of one or more end user utterances and the agent?s response or responses.

Important: A follow-up event generated outside the 24-hour period will generate a new engagement, and will not be included in the previous interaction. It will count as an individual event. The first agent message without a prior visitor message, such as a welcome message, doesn't create a new interaction.

What do the Interaction Trend and Interaction Type Trend charts show?

The Interaction Trend chart shows the total number of engagements per month as well as the deflection rate and the self-service score.

The Interaction Type Trend chart shows the number of AI-only engagements and the engagements that include both AI and agent live chat events.

What is the interaction self-service score and how is it calculated?

The interaction self-service score indicates whether the user was able to find a response to a specific request. It is calculated by dividing the number of interactions without negative signals with the total number of interactions.

An interaction is self-served if it meets the following criteria:

  • It doesn't include a channeling event
  • It doesn't include negative feedback
  • The bot was able to answer at least one user question

What data does the Interactions Summary table contain and how are they calculated?

The Interactions Summary table shows the following pieces of information about engagements:

  • Interactions: total number of interactions for the knowledge base
  • Bot interactions: total number of AI-only interactions for the knowledge base
  • Channeled: percentage of interactions that contain at least one channeling event like chat, ticket, or phone.
  • Channeled to agent: percentage of interactions that are channeled to an agent.
  • Agent only: percentage of interactions that include agent live chat only.
  • Self-service score: the percentage of interactions without negative signals divided by the total amount of interactions.

What time interval is used on this dashboard?

Interaction data for the last 12 months is available on this dashboard. The exact time interval is displayed under the Interactions Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated one a day. The last update time is displayed at the top, above the Interaction trend widget.

What do I see on the Interactions by Labels Dashboard?

The Interactions by Labels Dashboard provides an understanding of users? interactions by grouping them under the labels? hierarchies.

Note: The Interactions by Labels Dashboard is in beta. Contact your Customer Success Manager for more information.

The information on this dashboard is organized by the label hierarchy and by article. If you use labels to indicate business areas like products, services, you can find this dashboard useful to gain insight to users? experience as they relate to the different business areas. See Knowledge Base Labels to learn more about labels.

With this dashboard you can gain insights into the following:

  • Labels with the most interactions
  • Labels with the most channeled interactions
  • Labels with the most positive feedback
  • Labels with the most negative feedback
  • Other details on an article level such as the number of interactions, the number of resolved, unresolved and channeled interactions and the number of interactions with positive or negative feedback

You can do the following with the data on this dashboard:

  • See data for specific labels using the Label filter
  • Change the time interval to see a subset of the data
  • Select specific knowledge bases to see data for that knowledge base
  • Create bookmarks with the filters you set
  • View the individual widgets or the Label Summary table in full screen
  • Download the content of the dashboard, the individual widgets, or the Label Summary table to pdf
  • Hide columns and use drag-and-drop to change the order of columns in the Label Summary table

What is an interaction?

An interaction represents a user?s attempt to receive information or issue a request as it relates to a specific topic and includes all follow-up actions around the information retrieved. There are two types of interactions: AI-centric interactions and human-assisted interactions.

In an AI-centric interaction, a search event represents the start of an interaction and any follow-up events such as channeling, feedback and conversation events are included up to a period of 24 hours. An interaction conveys the treatment of a specific user need or intent, therefore it will include any piece of follow-up information that is needed to respond to the user's request.

A human-assisted or agent-facing interaction is comprised of one or more end user utterances and the agent?s response or responses.

Important: A follow-up event generated outside the 24-hour period will generate a new engagement, and will not be included in the previous interaction. It will count as an individual event. The first agent message without a prior visitor message, such as a welcome message, doesn't create a new interaction.

What data does the Label Summary table contain and how are they calculated?

The Label Summary table contains all the articles grouped by knowledge base and top 3 level of label hierarchy.

The Label Summary table shows the following pieces of information about articles:

  • Number of Interactions: number of interactions in the given knowledge base and article title.
  • Resolved: the number of interactions that are free from negative signals. Negative signals include:
    • Interactions that were not answered
    • Interactions that contain negative feedback
    • If multiple answers were provided by the bot but none got selected
  • Unresolved: the number of interactions that contain at least one negative signal
  • Channeled: the number of interactions that contain at least one channeling event like chat, ticket, or phone
  • Positive feedback: the number of interactions that contain at least one positive feedback event
  • Negative feedback: the number of interactions that contain at least one negative feedback event

What time interval is used on this dashboard?

Interaction data for the current month and the previous three months is available on this dashboard. The exact time interval is displayed under the Label Summary table. You can look at subsets of the data within this time interval by using the Date filter. By default, the dashboard displays engagement data for the last seven days.

How often is the data updated?

The data on this dashboard is updated one a day. The last update time is displayed at the top, above the Label filter.

How does Digital DX AI track conversion rates?

Note: Conversion tracking has been integrated into Acquire. For more information, contact your Customer Success Manager.

Digital DX AI responds to your users online and tracks the number of users who have successfully received answers through the AI Console before making a purchase.

For information on how to enable conversion tracking, see How can I add conversion tracking?.

What is the Answered Sessions chart?

The chart contains information about the number of answered sessions per each knowledge base as well as the total answered sessions in all knowledge bases combined.

To view the Answered Sessions chart, log into your Bold360 AI platform and go to Analytics > Traffic > Answered Sessions chart.

What is a Visit, Daily Unique Visitor, and Page load?

Daily Unique Visitor: Counts every visitor once in a time-frame of one day. Bold360 AI uses cookies for this tracking.

Visit: Visit is a time-framed period which user has visited a website. For every 30 minutes of inactivity, a new visit is counted.

Page Load: Counts for each time the widget loads.

How do I know how many monthly answers I have?

The answers shown in the Answered Sessions chart, which display how many sessions have been answered through Bold360 AI, including automatic answers, emails, and chats. To view the Answered Sessions chart, do the following:

  1. In the AI Console, go to Analytics > Traffic.
  2. Go to the Answered Sessions chart.

To download the chart, click on the download arrow located on the right corner at the top of the chart.

Note: If you have more than one knowledge base, the first chart (under 'My Account)' summarizes the answered sessions in all your knowledge bases. The chart below it (under ' X Knowledge Base statistics) summarizes only the answered session in the selected knowledge base.

Scenarios that are counted as answered sessions

What is the Q&A Traffic chart?

To view the Q&A traffic chart, in the Bold360 AI platform go to Analytics > Traffic.


Total Asked Questions
Shows you the number of total asked questions by visitors in the widget, and support center.
Answered by Bold360 AI
Shows you the number of automatic answers that were provided to the visitors by the widgets and support center.
Tickets
Shows you the number of escalations by email or forums that were performed in the widgets (Email button).
Chats
Shows you the number of chats escalations that were performed in the widgets (Chat button).
Click to call
Shows you the number of escalations by phone that were performed in the widgets (Phone button).

What is the Widget Traffic chart?

This chart shows an overview of your site's traffic on pages where Bold360 AI is visible for users. To view the chart, in Bold360 AI, go to Analytics > Traffic.

Go to the Widget traffic chart chart.

Page loads: Shows the number of page loads on pages where the Bold360 AI interface is implemented. In other words, a page load is when the widget loads.

Visits: Shows the number of visits to pages where the Bold360 AI interface is implemented.

Daily Unique Visitors: shows the number of new visitors on pages where the Bold360 AI interface is implemented.

What is the Widget Usage Rate chart?

This chart shows the percentage of your site's users who use Bold360 AI. To view the chart, go to Analytics > Traffic.

Go to the Widget Usage Rate chart.

Visit Usage Rate: The percent of interactions (page loads) where the customer used Bold360 AI to get an answer by typing a question or selecting an article.

Visitor Usage Rate: The percent of customers on the website who engaged with the widget.

What is the Escalation Reduction chart?


Escalation Reduction Rate
Questions that were not escalated to representatives.
Answered Rate
Questions that received an answer and were not escalated.
Escalation Rate
Questions that were asked and escalated to representatives.
Note: Escalation apply to email escalations (submitted via your contact form), while "Chats" apply to chat escalations.

To view the Escalation Reduction chart:

  1. Log into your Bold360 AI account.
  2. Go to Analytics > Traffic .
  3. Go to the Escalation Reduction chart.

Guidelines for optimizing your service with Search Optimizer (2.0)

Setting task priority

Priority should be set by balancing between the following voice characteristics:

Popularity
Displayed as Hits, which tracks the rate in which the voices are coming into the service, meaning the queries presented by your end users. Voices here come in handy since they provide you automatically with the clustering of queries to provide a straightforward insight on what is that current "hot topics".
The search optimizer categorization
  • Unanswered ? these are voices that the search engine did not answer
  • Answered ? these are voices that the search engine answered
  • Reactions ? articles that got direct feedback from end users. Reaction is a metric to reflect the ratio between positive and negative feedback. You can use this category to identify whether content is relevant and sufficient or needs update.
Trend

The trend shows the relative difference in the number of hits during the selected time interval and the time interval that precedes the selected time interval. So, if the selected time period is the current week, the trend is calculated relative to the previous week. When there is no value for a trend, it means that there were no hits either in the current period or the reference period of the trend.

Trend is calculated as follows:

For example, the selected time period is the current week and the number of hits is 500. The number of hits was 550 during the previous week. Using the above formula, the trend is -10%, so you?ll see 10% in red in the Trend column.

To recap, to prioritize your work, make sure that you look for the following:

  1. Popular Voices (high number of hits) that have not been answered (Unanswered)
  2. Popular Voices (high number of hits) that were answered but the answer was not satisfactory, indicated by either negative reaction or channeling
  3. Popular Voices (high number of hits) that the bot answered with one or more answers (Answered) to make sure that the answer provided is the right one.
Note: Depending on the size of your knowledge base and the number of customers you have, you may need a separate content manager to deal with answered voices.

Actions to be performed to handle a task

Fix a voice
For both unanswered and answered voices, you can either assign a voice to an existing article or create a new article. By selecting a voice in Search Optimizer, based on our NLP algorithm a list of relevant articles are displayed to which you can assign the selected voice.

When you select an Answered voice, the articles displayed on the right are those that were returned in a search widget (as opposed to a conversational one).

By assigning a voice to an article, you add that voice as a phrase for the selected article and subsequently, Bold360 AI will answer this voice and new voices that express the same intent with the assigned article.

If no relevant answer is suggested or you find that the intent is not covered by your knowledge base, a new article should be created by clicking Create new article.

Mute
When the voice reflects queries that you do not want to handle in your knowledge base, such a queries that contain gibberish, you can move them to the Muted category in order not to treat them at this point and focus on significant voices.
Delete
Once you have taken care of a voice in Search Optimizer, you can remove it from the list, so that you do not revisit the task that you have already done. Deleted voices are not moved to the Muted category, but are permanently removed from Search Optimizer. Deletion is used mainly for "housekeeping" - to clean up your Search Optimizer -, but it does not prevent your customers from asking the same question again.

Workload Management

Search Optimizer supports the collaboration of a team of content managers that work in parallel on the same knowledge base. You can assign an owner to voices and articles in Search Optimizer. By assigning an owner to voices, you can balance content optimization workload based on your business priorities and organizational structure. Each owner can then view only those tasks that are assigned to them and thus manage their work and thrive to empty their content optimization tasks in Search Optimizer.

To enable this workload management,

  1. Make sure that the Enable voices assignment feature is enabled for your account. Contact your Customer Success Manager to enable it for you.
  2. In the Bold360 AI platform, go to Admin Center > Users and set up the relevant users with the necessary privileges.

Prevent the bot from answering a question

In the Article Editor and in Search Optimizer 2.0, you can add Negative phrases, which tells the bots what voices an article should not answer. To do so, select an answered voice and click Do not answer on the right.

What are account health settings?

You use account health settings to control Dashboard sensitivity.

If you want to see more cases where your customers may not be getting the answers they need, set the threshold lower.

To change account health settings, go to Admin Center > KB Setup > Account health.

The following settings are available:

Name Description Default value
Health Threshold (%) Enables you to set the barrier between red and blue on the Dashboard. 0,5 which means that if 50% of the responses are negative, the voice or cluster will be red on the Dashboard. Similarly, if 50% of the responses are positive, the voice or cluster is blue on the Dashboard.
Minimum interactions with feedback Defined the number of interactions that are required before the voice or cluster is deemed red or blue. 2

What kind of data does traffic analytics provide?

Bold360 AI analytics provides set of predefined charts which shows amount of traffic coming through the system. These charts can be very helpful to determine tendencies in visitors interactions with widget and how it affects your website.

There are six charts in the Traffic section:

How can I view a list of unresolved questions?

Unresolved answers can be viewed on the Dashboard (as red tiles) or as a list.

The following user queries (Voices) are considered unresolved:

  • No answer was presented or selected
  • The question was escalated
  • The customer provided negative feedback
  1. Go to Analytics > Voices analytics.
  2. On the Voices analytics page, filter the list by Unanswered.
    Tip: The Hits column shows you how many times a voice went unresolved.

How can I remove personal data from Bold360 AI?

You can remove personal data manually from Bold360 AI.

Make sure you have modify permission for Analytics. If you don?t, you have the option to mark queries that in your opinion contain personal information. For more information see How can I report that I found personal data?.
In Bold360 AI users with view permission for Analytics have the option to mark user queries that contain personal data so that they can be removed. Users with modify permission for Analytics can review and remove the mark for personal data from the query or remove the queries that contain personal data.
  1. Go to Analytics > Reports.
  2. Run a report that contains the Query field.

    To review queries that users in the system marked for containing personal data, also select the Holds PII field and filter for queries that have True in this field.

  3. Find the query that contains sensitive data and choose from the following options:
    • Select (Sanitize Query) to obscure and remove the data.
      Note: This step is irreversible.
    • Click on (Personal Informaition in Query) to mark that the query doesn't contain personal data.

    Result: If you sanitized the data, the content of the selected query is replaced by asterisks (*) in the report and the query is removed from the Voices Dashboard as well as from Search Optimizer.
    If you marked that the query doesn?t contain personal data, the icon turns black and white and the value of the Holds PII column changes to False for the query.

What is Business Insights?

Business Insights enable you to analyze user journeys from beginning to end and gain insights into the efficiency of your digital engagement services.

Business Insights provides a comprehensive view of the user journey across the Bold360 AI and Bold360 Agent platforms: you can follow the end user?s journey from the moment they start interacting with the bot through their communication with the agent until the conversation ends.

Business Insights collects and presents data on the following dashboards:

  • Engagements Dashboard
  • Interactions Dashboard
  • Channeling Dashboard
  • Interactions by Labels Dashboard
  • Engagements History Dashboard
  • Interactions History Dashboard
Note: The Engagements Dashboard is available on all accounts. The rest of the dashboards are in beta. To learn more about the beta dashboards, contact your Customer Success Manager.

You can easily personalize the generic views of the Business Insights Dashboards to accommodate the your business needs and you can bookmark multiple personalized views for future use.

What event types are available in Bold360?

Event types are the most granular actions and occurrences detected on Bold360?s touchpoints.

Events are log entries of a user or system event that relate to the user experience. When you drill down in the Business Insights dashboards, you will see all the relevant engagements with their event types.

Bold360 AI event types

The following list provides information about event types that occur in Bold360 AI.

Answer link clicked
The user clicked a link in the body of the response to go to a linked Bold360 AI article.
Article collapsed
The user collapsed a response after viewing it.
Article expanded
When the search returned several results, the user expanded one of the options.
Autocomplete suggestion selected
The user selected an autocomplete option as they were typing their inquiry.
Channeled
The user requested to have their inquiry be handled by another channel, like phone or e-mail.
Feedback sent
The user provided positive or negative feedback.
Option selected
The user selected an option presented by the system, for example, a button that was part of a conversation or one response out of a few possible ones in a conversation interface.
Search
The user provided input on their need either by typing an inquiry or selecting from a list of predefined options, for example, in a support center or an FAQ list.

Bold360 Agent event types

The following list provides information about event types that occur in Bold360 Agent.

App chat created
The visitor clicked on the Chat button on an asynchronous channel.
App chat message closed after the timeout
The chat was closed after timeout on an asynchronous channel.
App chat message ended by agent
The chat with a live agent ended on an asynchronous channel.
App chat message has been escalated to an agent
The chat session was escalated to Bold side, directly to Automatic Chat Distribution (ACD) on an asynchronous channel.
App chat message received
The asynchronous text message between the user and the agent. The sender is indicated by the <<field>>.
App chat messages has been assigned to an agent
Asynchronous messaging chat was assigned to an agent by ACD event.
App chat started
Asynchronous messaging chat started.
Chat assigned by ACD
User was assigned to an agent by ACD. 
Chat assigned to agent
The user's chat was assigned to an agent, for example when ACD isn't turned on.
Chat button clicked
When the visitor clicked on the Chat button regardless of whether the button is available, unavailable or blocked.
Chat disconnected
The chat was disconnected.
Chat ended
The chat with a live agent ended.
Chat escalated
The chat session was escalated to an agent directly to ACD.
Chat message
Text message between the user and the agent. The sender is indicated by the <<field>>.
Chat queued
The user's request to speak with an agent was placed in the relevant queue.
Chat window closed
The user closed the Bold360 window.
Co-browse accepted
The user accepted the invite to co-browse.
Co-browse ended
The co-browse session ended.
Co-browse request
The agent sent a co-browse request.
Co-browse started
The user started to share their screen with the agent through a co-browse session.
Conversion
Specific chats can relate to actual sales or other milestones, such as new account created or page viewed. These events are tracked as conversions.
File downloaded
A file was downloaded to the user's device.
File sent
The agent sent a file to the user.
Pin invitation accepted
The visitor accepted the manual invitation.
Proactive invite accepted
The user accepted the not only proactive invitation.
Proactive invite declined
The user actively declined the not only proactive invitation.
Proactive invite ignored
The user ignored the event. It can be set up for any timeout.
Proactive invite sent
The user was identified as part of a target audience to be approached proactively. An invitation to engage was presented to this user.
Remote-control accepted
The visitor accepted the remote control request.
Remote-control ended
The remote control session ended.
Remote-control request
The agent started a remote control session.
Remote-control started
A remote control session started.

How can I retrieve the list of emails of users who escalated?

You can download the tickets report in excel spreadsheet. The report includes the list of emails.

To download the report, please follow the instructions below:

  1. Go to Analytics > SLA.
  2. Click Download Tickets Report and choose the relevant month:
  3. In the Excel file, look for the User email column to find the list of emails by users.

How can I track SLA for a certain label?

To track the SLA for a certain label (e.g. Quality Issue) you must enable SLA tracking for that label.

  1. In the AI Console, go to Ticketing > Labels, hover over the relevant label and click Edit.
  2. Under SLA select Track SLA Statistics for this label
  3. Schedule a custom tracking time, select Use custom SLA goals for this label and set the time according to your preference.
  4. Click Save.

Once you have the SLA tracking enabled, you'll start seeing data in the Label Statistics (under the Analytics > SLA. From that point you can click on the figure under the Tickets column and select the relevant target date.

How do I set the SLA tracking?

SLA (Service Level Agreement) Tracking allows you to get a clear picture of your support performance and activity.

Tickets that are waiting for longer than the SLA period defined will be marked as 'Overdue' in your inbox.

  1. In the AI Console, go to Analytics > SLA.
  2. Under Tickets SLA Tracking, click SLA Settings .
  3. Set the following SLA Tracking conditions:
    Name Description
    Ticket SLA Resolution Time SLA for closing tickets
    Ticket SLA Response Time SLA for replying to incoming tickets.
    Business Hours Tracking Determine the working days and hours on which the SLA will be tracked. You can chose between calculating the service availability by your service center working hours (specific days and hours or 24/7). You can also define specific days (as holidays) to be excluded from the calculation.

How do I measure Bold360 AI business value?

Depending on your business goals, there are numerous KPIs for tracking the benefits of your Bold360 AI implementation.

KPIs can be prioritized, baselined, and tracked across time. Some KPIs are tracked within Bold360 AI, while others can be extracted from your complementary business operations systems.

Here are some KPIs you may want to consider:

Category KPI Comment
Self-service centricity & success Percentage of self service interactions out of all interactions  
Self-service centricity & success Percentage of human assisted contacts originating from self service  
Self-service centricity & success Percentage of self service success rate Bold360 AI dashboard
Assisted channels Chat volume Normalized: visits/customers
Assisted channels Inbound Phone Calls Normalized: visits/customers
Assisted channels Emails and tickets Normalized: visits/customers
Workforce effectiveness Amount of emails and tickets  
Workforce effectiveness AHT (average handle time)  
Workforce effectiveness SLA compliance  
Workforce effectiveness Average ticket resolution time (elapsed)  
Customer Experience Customer effort score (CES)  
Customer Experience Net promoter score (NPS)  
Customer Experience Customer satisfaction (CSAT)  
Customer Experience Channeling accuracy (e.g. from LP) Also for WF effectiveness
Business results Sales conversion rate  
Business results Churn/attrition  

How do I view statistics of 'liked' answers?

  • To view statistics of liked answers, navigate to Analytics > Custom Charts > Usage > Like Rate.
  • You can also view the number of liked and disliked answers per each Q&A by selecting Knowledge > All articles.
Example:

How can I create an Organic Traffic chart?

  1. Go to Analytics > Custom charts.
  2. Go to the Support Center tab with the parameters that are of interest.
  3. Choose the Organic visit check box
  4. Click Create Chart.
  5. In the top-right corner of the chart, you can define the date range for your report and display an updated chart.

Tickets SLA Tracking

SLA charts enable tracking your ticketing handling performance.

  • The Response Time and Resolution Time charts show you information regarding tickets first response average time (Response Time) as opposed to tickets closing average time (Resolution Time).
  • The Tickets Status Overview list displays the values of different ticketing statuses, including: replies, open, closed and overdue tickets for last week. Please click on one of the metrics to get a chart for daily/weekly/monthly data.
  • The Goal Overview pie chart displays the percentage of Good SLA, Fair SLA and Overdue according to your goal settings.
  • The Label Statistics chart displays the values of different ticketing parameters for the defined period, sorted by custom labels such as sales or urgent. Click here to find out how to define the labels you want to track.
  • The Agent Performance chart displays the values of different ticketing statuses, including: replies, open, closed and overdue tickets for the defined period, per agent.
    Note: Data is displayed according to the SLA settings in your account. You can change it by clicking Settings.

See How do I set tracking?

Export Tickets' SLA Information

To export tickets' SLA information, do the following:

  1. Go to Analytics > SLA.
  2. On the upper right side of the screen, click on Download Tickets Report..
  3. Select the month and year to export. You can also select a label and choose to include ticket correspondence (both are optional).
  4. Click Download Report
    Note: The data is exported in XML/ CSV format.

Where can I see an article's feedback?

You can find customer feedback at several places in Bold360 AI. For information on gathering customer feedback, see this article.

Dashboard > Feedback tab
On the Dashboard > Feedback tab you can see voices with negative and positive feedback.
Reports
On the Analytics > Reports page, run a report with Query, Feedback Type, and Feedback Reason selected to see all customer feedback.
API
You can also use the following API call to automatically generate the same information as you see in the Reports menu: https://my.nanorep.com/api/doc/v1.html#q=Get Reports

What are the custom charts?

Custom charts provide visual representation of a variety of parameters and allow you to explore their behavior over a selected period of time.

To create a custom chart, do the following:

  1. In the AI Console, go to Analytics > Custom charts.
  2. Click on a tab to select a category that you are interested in.
  3. Select the fields you would like display on a chart.
  4. Click Create Chart to see the result.
  5. Select the date range and the level of details from the Analysis drop-down list in the top-right corner of the chart to display an updated report.

You can Download the data by clicking the download icon under the date range selector.

What charts are available?

You can display the following custom charts:

Traffic page

Page Loads: The number of times visitors loaded a page that contains a Bold360 AI widget.

Visits: The number of viewing sessions by visitors to your site. A visit starts when a visitor lands on your website. The end of the visit typically occurs when the visitor leaves your site or is inactive for more than 30 minutes. The difference between page loads and visits is that a visit will only be counted once per session. Meaning if a user opens two pages both containing a Bold360 AI widget that would count as two page loads but a single visit.

Engagements: A visit in which Bold360 AI capability was used at least once (such as typing a question or selecting an article) in order to fulfill their need. An engagement may include one or more interactions.

Non-Active Visits: The number of visits in which a visitor did not engage with the Bold360 AI widget. Visitors had the option to use the widget but did not necessarily use it.

Absolute Unique Visitors: The number of unique visitors who visited a page that contains the Bold360 AI widget.

Absolute Unique Active Visitors: The number of unique visitors who visited a page that contains the Bold360 AI widget and used it.

Returning Visitors: The number of Absolute Unique Visitors who had seen the widget more than once, but did not necessarily used it.

Daily Visitors: The number of visitors each day on a page that contains the Bold360 AI widget. This number is the total of Daily Active Visitors + Daily Non-Active Visitors.

Daily Active Visitors: The number of visitors who actively engaged with the Bold360 AI widget on a page each day.

Daily Non-Active Visitors: The number of visitors who visited a page that contains the Bold360 AI widget but did not engage with the widget.

Total Asked Questions: The number of interactions that you can also view on the Dashboard. Interactions include answered, unanswered, and escalated questions.

Answered by Bold360ai (including escalated): The number of customer questions that were answered, including those that were also escalated. This number is the total of Total Resolved by FAQs + Total Resolved by support center + Total Resolved by Search + Total Answered by Auto-Complete.

Total Answered and not Escalated Questions: The number of customer questions that were answered and not escalated.

Total Resolved by FAQs: The number of questions that were answered by FAQs, which you can set up in the touchpoint configuration for widgets and Support Centers as well.

Total Resolved by support center: The number of questions that were answered by the navigation are on the Support Center.

Total Resolved by Search: The number of questions that were answered by the search widget.

Total Answered by Auto-Complete: The number of questions that were answered by the user selecting an auto-complete suggestion.

Liked: The number of answers that customers liked by clicking the appropriate button.

Disliked: The number of answers that customers disliked by clicking the appropriate button.

Total Escalations: The total number of questions that customers escalated. This number is the total of Chat Escalations + Chat Invitation Escalations + High-Value Chats + Email Escalations + Click To Call Escalations

Chat Escalations: The number of questions that customers escalated through chat.

Chat Invitation Escalations: The number of questions that customers escalated in a chat invitation.

High-Value Chats: The number of questions that were considered as high-value and which automatically kicked off a chat escalation.

Email Escalations: The number of questions that customers escalated through email.

Click To Call Escalations: The number of questions that were escalated by phone (by clicking the Phone button).

Usage

These charts provide information about the usage of Bold360 AI widgets.

Visits Usage Rate: The percentage of visits during which the customer engaged with the Bold360 AI widget.

Visitors Usage Rate: The percent of visitors who use the Bold360 AI widget.

Escalation Reduction Rate: The percentage of customer questions that were not escalated.

Like Rate: The percentage of answers that customers liked by clicking the appropriate button.

Escalation Rate: The percentage of customer questions escalated to any channel.

Satisfaction Rate: The percentage of customers who were satisfied with the response they received. Calculated as Answered questions with no negative signal (not channeled and no negative feedback given) / number of questions.

High-Value Chat Rate: The percentage of high-value questions that automatically kicked off a chat escalation.

Support Center

These charts provide information about the usage of your support center.

Visits: The number of customers who visited a page that contains the Bold360 AI widget.

Page loads: The number of pages containing the Bold360 AI widget that were visited.

Organic Visits: The number of visits generated via search engines, such as Google.

Total Resolved by support center: The number of questions that were answered by a Support Center.

Knowledge-Base Size

These charts provide information about the content of your selected knowledge base.

Questions in Inbox: The number of tickets that you have received in the Bold360 AI ticketing system if the ticketing feature is available in the knowledge base.

Ready to Publish: The number of articles that your agents suggested as new content and have not been published yet. You find these articles on the Knowledge > Suggested Content page.

Online Questions: The total amount of phrases (including titles) modeled in the knowledge base.

Online Answers: The total number of articles that you have in your selected knowledge base.

Knowledge-Base Usage

These charts provide information about the tickets in your selected knowledge base.

Answered Tickets: The number of tickets answered in the selected knowledge base.

Added Tickets: The number of articles added to the selected knowledge base.

Deleted Tickets: The number of articles removed from the selected knowledge base.

Users

These charts provide information on each user of your Bold360 AI account.

Online Answers Added: The number of articles added to the selected knowledge base.

Deleted Answers: The number of articles deleted from the selected knowledge base.

How to differentiate between positive and regular channeling for high-value chats

Positive channeling indicates intentional channeling. For example, when a specific intent indicates a customer is at risk and your strategy is to help this customer with a live agent. To add channeling value to an article, select High value for Channel Value on the Options tab in the Article Editor.

This impacts channeling statistics between high-value and regular channeling on the Dashboard.

Important: Channeled high-value interactions are still considered as unresolved from a self-service point of view. Therefore, the self-service score will be 0 for an interaction that was channeled from a high value article.

Why do I see different figures in the exported SLA data (compared to the dashboard data)?

The exported SLA data is based on UTC+0. The dashboard SLA data is based on the viewer's time-zone.

Announcements

Bold360 21.10 Release Notes

Bold360 21.10 release notes

Bold360 version 21.10 contains the following new enhancements:

Enforcing HTTPS and removing HTTP across widget scripts

Going forward, there will no longer be an option to choose (non-secure) HTTP-only connections when generating widget touchpoint scripts. HTTPS will be enforced across all touchpoints going forward and HTTP-only widget configurations will no longer be supported as of December 31, 2021. Any widget script using HTTP only will need to be changed by this date in order to continue uninterrupted

Access to the Genesys DX AI management (*.nanorep.co) will only be accessible through HTTPs

In-product announcements will be present to raise awareness of this change

Important: Any widget script using HTTP only will need to be changed by this date in order to continue uninterrupted.

In-Product Rebranding

As part of the transformation to Genesys DX, the product interfaces are now rebranded to reflect this change. Logos, product names, and company links are now updated to Genesys branding and is present within each of the product interfaces.

getEstimatedWaitTime API call

Using the "getEstimatedWaitTime" API call will provide the estimated wait time that the visitor would spend waiting before starting a live chat with an agent.

Details on using this API can be located in this article.

Bold360 21.11 Release Notes

Bold360 version 21.11 contains the following new enhancements:

New KB Launchpad Intent Library for COVID-19/Vaccinations

In order to help businesses with providing just-in-time information related to updates to COVID-19 and vaccination protocols, a new entry in the KB Launchpad is now available with intents that can be used to deploy knowledge to keep your workforce informed of new policies and information that you may need to enforce.

Not familiar with the KB Launchpad? The KB Launchpad contains pre-defined intent libraries to help you add industry and use-case specific questions/intents to your KBs with your own specific responses to help with adding commonly asked questions for newly created KBs.

We will be enabling the KB Launchpad as a default setup option on all accounts shortly but contact your Account or Success Manager to get it enabled early if necessary.

ICO Cookie Compliance

In order to support ICO cookie compliance in regions that require end-user consent for allowing cookies to be set, Genesys DX now complies with this standard by making the tracking and analytics related cookies that are set through the touchpoint and visitor monitoring HTML snippets to be optional.

By default, all cookies will continue to be set upon page load. To restrict these cookies from loading automatically and only upon end-user consent, additional steps will be required on the website that the Genesys DX code is being deployed to.

You can learn more about what this compliance is and how to update your Genesys DX deployment to be compliant by reviewing the details in this article: ICO Cookie Compliance.