Discover: Getting Started Guide

Forethought Support

Updated June 13, 2024 14:38

Discover automatically extracts granular insight from your customer support tickets. It leverages advances in Natural Language Understanding to cluster tickets together into topics.

To view Discover, click on the Discover tab on the left menu.

Clicking Discover should get you to the dashboard view.

Dashboard

Time selector

Time selector sets the same time frame to provide insights into customer support ticket data.

Default is 30 days.

Other options for time selector include:

Select one of them to see KPI's for customer success in the selected time period.

Time analysis selector

Daily	Provides graphs to represent aggregate data for all tickets on a daily basis.
Weekly	Provides graphs to represent aggregate data for all tickets on a weekly basis.
Monthly	Provides graphs to represent aggregate data for all tickets on a monthly basis.

Selecting one of these lets you hover over any of the graphs for each metric and compare it to metrics in the previous interval.

E.g. when you select daily and last 30 days today’s metric will be compared with the metrics from 31 days prior.

Customer Support KPI Metrics and Definitions

Metrics

Volume
First resolution time
Full resolution time
Number of agent replies
First contact resolution (%)
Reply time
Sentiment

Definitions

Metric/Section	Description	Measurement Unit	Example
Most common topics	Topics with the highest volume over the selected time period
Top movers	Topics whose metrics have changed the most over the selected time period versus the previous time period
Volume	Total number of tickets in the given time period	Number	873 tickets
First resolution time	Average time taken to resolve a ticket for the first time	Hours	4 hours
Full resolution time	Average time taken to resolve a ticket for the last time	Hours	4.2 hours
Number of agent replies	Average number of times an agent responded to a ticket	Number	3 replies
First contact resolution (%)	Average percent of time a ticket is resolved after the first contact	Percentage	20.5%
Reply time	Average time taken for an agent to respond to a ticket	Hours	2.6 hours
Sentiment	Sentiment is automatically detected for the starting and ending sentiment.	0 means negative sentiment 50 means neutral sentiment 100 means positive sentiment	73
Starting sentiment	Ticket sentiment when the ticket was created.	0 means negative sentiment 50 means neutral sentiment 100 means positive sentiment
Ending sentiment	Ticket sentiment on the last reply from the user before it was resolved/closed. This is same as starting sentiment when there is only one user query and no replies.	0 means negative sentiment 50 means neutral sentiment 100 means positive sentiment
Drop in sentiment	Number of tickets which ended with lower sentiment from where it started	Number	400
Percentage change	Performance comparison of a topic between the current selected time period versus the previous time period. Positive (+) percent change indicates the metric has increased, while negative (-) percent change indicates the metric has decreased compared to the previous period.	Percentage
Deviance	Comparison between one topic’s metric to the average of all topics. Positive (+) deviance indicates the selected topic’s metric is more than the average, while negative (-) deviance indicates the selected topic’s metric is less than the average.		If a topic’s volume deviance is +2.3x, the topics has 2.3 times more tickets than the average. If a topic’s volume deviance is -1.2x, that topic has 1.2 times fewer tickets than the average.
Channel	Means of communication that your customer uses for their support queries.		Email, Text, Voice etc.
ID	Unique ticket identifier in your help desk	Alphanumeric	September 8, 2022 11:00 AM
Date created	Date when the ticket was created	Date
Resolved on first contact	Was it resolved after the first reply?	Yes/No
Subject	Ticket subject in the help desk
Body	Ticket body in the help desk

Top Movers

This section shows those topics that had maximum change over the previous selected time period. The change can be positive or negative. % change in green indicates positive change and in red indicates negative change.

Most common topics

This is the list of the top common topics by volume.

Bookmarked Topics

This section provides insights on topics you easily wan to watch/follow with their respective metrics

Notifications

At the top right an an icon for notifications shows if you have any new notifications.

All Topics

In this view, you see auto generated topics that cluster tickets together. You can select the time period you want to see the topics and the associated metrics for.

How are topics automatically generated?
We use the advances in NLP and NLU to understand and comprehend the ticket’s subject and body and cluster them together into  auto generated topics that best represents the cluster of tickets.

In addition, you can select the metric you want to focus on.

Here you have two additional columns

Percentage change in metric
- This column show how the average metric changed from the selected time period to the previous time period.
  - Example → In the image below you can see that the topic return request changed by +40% compared to the previous 30 days.
Deviance metric
- This column shows if a metric is behaving in an abnormal way compared to the other topics.
  - Example: The +4.0x for return label request means that first resolution time for this topic is 4 times more than the median across all topics.

Only sentiment has different values

Starting sentiment
Ending sentiment
Tickets that dropped in sentiment

In addition you can also filter the results based on ticket and agent meta data.

This dashboard is refreshed daily.

Upon clicking on a topic you can go into the detailed view of the topic.

Topic View

Here you can select a specific metric to focus on. You can also select the time period and the time range you want to compare with.

You can also see the recommendation and the savings automating this topic into a Solve workflow can bring.

You can also see tickets that were part of the selected topic in sample tickets below the metrics graphs.

You can also click on Generated articles to see the articles Forethought AI generated for this topic.

You can edit the topic name by clicking the pen icon next to the topic name.

Note: Currently only 250 tickets randomly selected in the time period are shown.

Ticket View

Clicking on a ticket shows the subject and body of the ticket. It also shows the ticket ID, the ticket channel, when it was created, was it resolved on first contact and its metrics.

Clicking on the ticket id takes you to your help desk to view the ticket directly.

Icons and what they mean

Icon	Description
	Easy way to tag topics for quick access
	Discover recommended automations for workflows for better self-serve support
	The recommendation for the topic has been automated and now has a workflow under Solve workflows.
	Any topic with this icon means a new topic has been identified and the topic taxonomy has been updated.
	Topics that have been hidden from the view.
	This means, this topic has Knowledge base article gap and Forethought AI has generated articles.
	Forethought AI has generated one or more article.

FAQ

Question	Answer
How many tickets do I require for Discover to create a ticket cluster taxonomy?	You need a minimum of 100,000 tickets.
How many tickets do I require daily for Discover to cluster tickets in topics going forward?	Once we have 100,000 historical tickets, we can cluster as little as 1 ticket a day going forward.
How often is the dashboard updated?	The dashboard is updated daily.
Can I change the topics?	The topics are currently automatically generated and but you do have the option to edit the topic name in the topic view. If you have any questions, please reach out to support.
How do I get started?	All we need is to connect with your help desk and get access to your historical tickets. We currently have multiple help desks you can connect very easily to. To know more about our integrations visit integrations page.