Freshworks Analytics offers AI-powered capabilities through Freddy AI Insights, which are available to all the Pro and Enterprise users with a Freddy Co-pilot license, and currently available as early access for the Freshchat, Customer Service Suite, and Freshservice products.
To access Freddy AI Insights:
1. Go to Freddy > Insights.
2. Select the relevant insight.
Note: Freddy AI Insights is currently in public beta.
This article contains the following topics:
- The power of Root Cause Analysis: Moving beyond descriptive insights
- Benefits of Root Cause Analysis
- Ways of conducting Root Cause Analysis
- Conclusion
- Frequently asked questions
The power of Root Cause Analysis: Moving beyond descriptive insights
Root Cause Analysis (RCA) helps organizations move beyond surface-level insights to uncover actionable and fundamental factors that can be addressed to enhance customer and employee experiences, and improve business outcomes.
Benefits of Root Cause Analysis
1. Identify the underlying factors
By pinpointing the fundamental root causes, organizations can address the real issues rather than just the symptoms.
2. Understand causality
RCA helps determine these relationships, allowing organizations to identify the most influential factors and prioritize them for action.
3. Enable problem-solving
RCA provides the basis for developing targeted strategies, interventions, or solutions to address underlying problems or leverage positive factors driving desirable outcomes.
4. Improve decision-making
RCA enhances data-driven decision-making processes by offering a comprehensive understanding of the factors contributing to specific insights or outcomes.
5. Enable proactive measures
Organizations can be equipped to take proactive measures to prevent the likelihood of similar issues by enabling preventive actions, process improvements, or targeted interventions based on RCA insights.
Ways of conducting Root Cause Analysis
Dimension Deep Dive
Analyze the same metric through different dimensions to identify which ones contribute to the observed changes. This method helps pinpoint the top contributing dimensions without the need for extensive slicing and dicing of data.
Example Insight: "CSAT ratings decreased by 32% on 20th May 2024."
This approach quickly reveals which dimensions are most impacted, allowing targeted actions and strategies to address specific issues or capitalize on opportunities.
1. Region: CSAT ratings decreased by 28% in the US region on 20th May 2024.
2. Priority: CSAT ratings decreased by 21% for high-priority tickets on 20th May 2024.
3. Channel: CSAT ratings decreased by 30% for email tickets on 20th May 2024.
Enabling Root Cause Analysis
Whenever a user accesses the Freddy module within the product, they will see a list of all the insights that are proactively generated. These insights are user-role based, therefore only relevant insights will be shown to each user.
Users just need to select any insight to get a deep-dive into the leading root cause.
The root causes appear in an intuitive tree chart representation, with each node (box) representing data.
There is also a summary in natural language, which simplifies the interpretation for users.
Users can also take a deeper dive into all the other contributing factors with the “View all causes” option.
Example: Enabling Root Cause Analysis
Users access the Freddy module and discover proactive insights on the left bar.
On selecting the insight, the user will see a chart, which visually depicts the insight trend.
Following the trend graph is the Root Cause Analysis in the form of a summary in simple language followed by a tree chart, which explains the leading cause behind the insight.
Users can further drill down and take a look at the other underlying causes with the “View all causes” option.
Explanation of the example in detail
In the above example, the insight is about a surge in resolution time compared to last month.
The first node (box) of the tree chart shows the Average Resolution Time insight, which has gone up from 20 hours last month to 60 hours this month, an increase by 40 hours or 200%.
Similarly, it also shows that the Total Ticket Count in the first node has gone up to 4200 tickets, an increase by 1456 tickets compared to last month.
As we examine the linkages in the connected nodes, the surge in the resolution time is related to the Cloud Security Group. Compared to last month, there has been an increase of 60 hours in the Average Resolution Time and an increase of 1680 in total tickets.
On further examining the Cloud Security Group, the increase in Average Resolution Time is seen in two categories:
Hardware Category
Network Category
Hardware Category
For the total of 1460 Tickets belonging to Cloud Security and Hardware Category, the Average Resolution Time is 120 hours, an increase of 60 hours from last month. Most of the tickets belonged to Urgent and Medium priority where the resolution time has increased.
For the Urgent priority tickets of Hardware Category, the First Response Time was breached for 10 Tickets and the Average Resolution Time for those tickets is 300 hours, which is a 200 hour increase compared to the last month.
Network Category
For the total of 280 Tickets belonging to Cloud Security and Network Category, the Average Resolution Time is 20 hours, an increase of 15 hours from last month.
Additionally, the First Response Time is more than 40 hours for 4 tickets, an increase of 2 tickets compared to last month and the Average Resolution Time for those tickets is 190 hours, which is a 100 hour increase compared to the last month.
Conclusion
Root Cause Analysis is a powerful tool for moving beyond surface-level data insights. By identifying the underlying factors, understanding causality, enabling problem-solving, improving decision-making, and facilitating proactive measures, RCA empowers organizations to make meaningful improvements and drive better business outcomes. Using methods like the Dimension Deep Dive ensures a focused and efficient approach to uncovering the true causes behind observed data trends.
Implementing RCA in your organization will help solve current issues and pave the way for continuous improvement and long-term success.
Frequently asked questions
1) What are the frequently used terms related to RCA?
Tree chart: The visual representation for showcasing the interlinkage of causes.
Nodes: Boxes in the tree chart showcasing data, representing the underlying causes.
Critical Path: The leading/most probable underlying root cause.
Summary: A consolidated summary of all the data in different nodes of the tree chart, in simple language.
2) Where can I view RCA?
RCA can be accessed from the Freddy module in the product.
3) Is RCA applicable for all types of insights?
RCA is generated for only Proactive Insights and is not applicable for charts generated through Conversational Analytics.
4) What timeframe is used to create the RCA for an insight?
This depends on the type of insight as follows:
5) What is the refresh frequency for RCA?
The refresh rates for RCA depends on the refresh frequency of Proactive Insights. It follows the following pattern:
Note: RCA is pre-computed and the data point may vary with current number (based on the above refresh frequency).
6) What types of fields and metrics is RCA currently applicable for?
7) What do the terms “unknown” and “others” mean in RCA?
Unknown: It refers to the empty value in any field.
Others: This refers to the multiple values in the data of any field without any significant portion from a particular value.
8) What all tickets are part of the leading cause underlying data?
All the tickets present in the leaf node (first node) of the critical path of the tree chart.
9) What all tickets are part of the view all causes underlying data?
All the tickets of that day/week/month for which the insight is generated.
10) Can I add or edit any particular field in the treemap for RCA?
No, currently only the fields identified by the AI-model are part of the tree chart.
11) Is RCA available for all insights?
No, RCA is currently not available for all insights.
12) Does RCA support multi-language?
No, currently it does not support multi-language, but it’s being considered for future roadmap.
Related links:
Gain visibility into service operations with Generative AI-powered Freddy insights.
