Customer Experience

Speech Analytics for Call Centers: 5 Use Cases & Tools

Speech Analytics for Call Centers: 5 Use Cases & Tools
Customer Service Researcher
LinkedIn icon
Speech Analytics for Call Centers: 5 Use Cases & Tools

Call centers get thousands of phone calls a day, but only 2% of them turn into useful insights.

The other 98%? They go into a black box. 

Historically, it’s so difficult to extract insights from speech because of the sheer volume and the qualitative nature of it.

Call center leaders would either:

  1. Manually listen to a sample of the calls and tag each with keywords. This is time-consuming and not comprehensive enough. You also don’t have confidence in the evidence to send it to other teams to push for changes. 
  1. Do a huge diligent analysis once a year. This takes a technical skill set to do AND to understand the output. By the time you get insights, they’re already obsolete. 

So, what’s the solution?

Enter: Speech analytics.

In this article, we’ll go into what speech analytics is, how it works, and how you can use these insights to improve customer experience and call center performance.

Jump to:

TL;DR

1. Speech analytics for call centers is the process of analyzing recorded conversations between agents and customers using AI to extract useful insights.

2. Call centers use speech analytics to identify areas of friction in the customer journey, improve CSAT scores, assess agent performance and ensure high call quality, optimize the product roadmap, and get buy-in from other departments.

3. When looking for a speech analytics tool, prioritize one that analyzes all feedback channels, offers granular insights, is easy to use, and comes with integration support.

4. The three best speech analytics software in 2024 are SentiSum, Calabrio, and NICE CXone.

What is Speech Analytics for Call Centers?

Speech analytics for call centers is the process of analyzing support calls at scale using AI. 

AI models like Natural Language Processing listens to each call (much like a human would) and labels that conversation with a predetermined tagging taxonomy.

This includes:

  • The main topic of the call
  • Customer sentiment (e.g. Happy, sad, angry)
  • Priority (e.g. Urgent, not urgent)

For instance, let’s say 100 customers contacted you about damaged packaging during the delivery process. AI listens to, transcribes, and categorizes these calls and gives you quantifiable insights, such as:

  1. Topic: Damaged Packaging During Delivery
  2. Sentiment: 94% Negative
  3. Priority: Urgent

On a speech analytics platform like SentiSum, you can also see whether these issues are increasing or decreasing over time. If, for example, 110 customers are now contacting you about the same issue, you can quickly take action before it becomes 200 or 300. 

Source: SentiSum’s Speech Analytics Dashboard

Here are some other ways a speech analytics platform helps call center leaders:

  1. Understand agent performance at a glance. A speech analytics tool transcribes each individual support call so you can understand how each agent handled a particular situation. This information is crucial to build agent training modules and improve call quality as a whole. 
  2. Analyze support conversations from all channels. Most speech analytics tools bring in data from all your support channels, such as emails, chats, CSAT/NPS surveys, reviews, and more. This helps you understand the main issues customers are contacting you across all your channels and the sentiment behind them.
  3. Uncover the main drivers behind CSAT scores. Speech analytics software helps correlate CSAT scores with past interactions on all channels. This gives more context to low or high scores.
  4. Build evidence-based cases for change. Speech analytics platforms turn qualitative calls into quantitative data at scale. This data is crucial for call center leaders to push for improvements in other departments and answer ad-hoc questions they may have.
Companies that leverage speech analytics saved 20% to 30% in cost, improved CSAT scores by 10% or more, and saw stronger sales. Source: McKinsey & Company

So, how does a speech analytics tool work exactly?

How Call Center Voice Analytics Works: The Tech Behind

Call center voice analytics tools use AI-powered natural language processing (NLP) to analyze large volumes of conversations between agents and customers. 

In a nutshell, here’s how it works:

  • First, the software records each support call.
  • It then uses speech recognition technology to transcribe the spoken words into text.
  • A NLP model then analyzes the transcribed text, uncovering key topics and sentiment.

To give you a full picture of how speech analytics software works, we’ll use our own tool–SentiSum– to illustrate the process. 

(For the visual learners, here’s a 30-second YouTube video of a summary breakdown. We recommend watching it at 0.5x speed. ⬇)

Part 1. Integration with your existing help desk platforms. 

The speech analytics software first brings in voice call data from your helpdesk platforms. For example, SentiSum integrates with most major (and minor) helpdesk platforms like:

  • Zendesk
  • Freshdesk
  • Gorgias 

You can explore the full integration list here

But apart from voice calls, SentiSum also brings in data from all other feedback channels, such as surveys, customer support chats, emails, and more, onto one dashboard.

Part 2. AI analyzes voice calls. 

Instead of listening to calls manually yourself or having your agents add tags, SentiSum automatically analyzes and applies detailed tags using machine learning-based artificial intelligence.

In the example below, you can see how powerful machine learning NLP is compared to keyword extraction (what platforms like Zendesk typically use) and rule-based NLP (what other software providers use).

Learn more about different types of NLPs here.

Different types of NLP models - SentiSum
Different types of NLP models

Part 3. See customer call topics and trends. 

Once the analysis and tagging process is done (usually takes seconds), SentiSum’s easy to use dashboard shows all the key insights, like:

  • Trend of voice call volume over time 
  • Main issues for contact (we call them “topics”)
  • Top increases and decreases in sentiment and volume
A screenshot of SentiSum's insights dashboard - SentiSum
SentiSum's insights dashboard

If you click into each topic (e.g. “Request Refund”), you can see:

  • All conversations tagged under the topic
  • Increases and decreases over time
  • AI chatbot called “Dig In” - which allows you to ask questions and get succinct, summarized answers.
A screenshot of SentiSum's dashboard - SentiSum
Detailed breakdown each topic

Want to dig into how each support call went? No problem.

You can click into each call conversation and get:

  • An AI summary of each call 
  • Agent performance analysis
  • Transcript of the call

Part 5. Insights are pushed to your help desk.

These insights shouldn’t just sit in your analysis tool, but also be pushed back to your customer facing platform. 

On SentiSum, you can create automation workflows using a custom AI model and push triage and prioritization rules back to your helpdesk platform. 

This helps your agents understand which issues to fix first based on sentiment and urgency (e.g. angry customers need to be attended to quickly).

Want to see what the platform looks like and how it can help you? Please book a demo below!

SentiSum's CTA

How Do Call Centers Use Speech Analytics? 5 Use Cases to Improve Customer Experience and Agent Performance

So, you’ve got the insights. Now what? 

Here are five ways you can use the insights from speech analytics to improve your product, customer experience, and teams. 

Use Case #1: Identify Friction in the Customer Journey

A customer calling your service team represents friction. Whether it’s a delivery gone wrong or a damaged product, these are issues customers can’t fix themselves.

These issues, if they become a bad-enough experience, can cause customers to churn and choose competitors in the future, leading to revenue loss.

The key is to identify these friction points early on before they escalate. This way you can draw up plans to fix the issues and improve customer experience, driving loyalty and revenue in the future.

On SentiSum, you can see the main topics that are decreasing or increasing in volume over time (daily, weekly, or monthly).

This helps you quickly understand (and quantify) which issues you should prioritize fixing and use this data to push for improvements in other departments. 

For example, Lakrids, a gourmet D2C company, used SentiSum to understand what customers are saying across their Dixa support tickets. 

They noticed that towards the end of 2021, there was an influx of tickets regarding damaged packaging during delivery.

Based on this insight, their Head of Customer Service and Sales Support drew up a plan to improve packaging and transportation, along with a compensation plan for affected customers.

This initiative led to a 26% decrease in packaging-related complaints within a year.

Løkke Engraf, Head of Consumer Service and Sales Support @ Lakrids

Use Case #2: Improve your CSAT scores

Call centers usually send out CSAT surveys at the end of each support conversation.

But CSAT scores mean very little without context. Without being able to correlate CSAT data with your support tickets and agent conversations, you have to rely on hunches to figure out what went wrong.

SentiSum maps your CSAT scores to your voice call topics automatically, so you understand why a customer gave you a low (or high) score even if they didn’t leave a comment. This is key to fixing issues at the very root and improving CSAT scores in the future.

You can also use these explanations to coach your agents to provide better customer service. 

🎙️Related podcast: How to drive customer service CSAT from 30% to 90% 

Going back to Lakrids, our platform helped make sense of the vast amount of disparate customer data they had in their support platform, Dixa. 

By connecting each CSAT score with matching support tickets, we helped them understand why customers left the feedback they did. 

In less than 12 months, Lakrids improved their CSAT scores by 9%. 

Løkke Engraf, Head of Consumer Service and Sales Support @ Lakrids

Use Case #3: Ensure agent quality.

Training agents is hard when you don’t have the full picture of what’s going on.

If you’re randomly analyzing a small section of agent calls, you’re likely to miss out on broader trends of agent performance. 

Speech analytics can help you implement tighter standards for quality assurance by analyzing ALL your support calls.

For instance, SentiSum shows you the transcripts of each call under each topic. You can click into each to understand what went wrong and what went well. 

Our ChatGPT-like AI also gives you a summary of how each call went, so you don’t need to read the entire conversation. We also recently launched ‘Analysis’ (Beta) which gives you a breakdown of agent performance. 

Together, they provide detailed feedback on how specific agents handle each interaction. For example:

  • Was the issue resolved? How long did it take to be resolved?
  • Was the agent polite and empathetic towards the customer?
  • Was the agent professional in handling the call? What could they have done better?

You can then use these insights to set up agent training modules or to improve a particular agent’s performance. 

For instance, Scandinavian Biolabs, a haircare D2C brand, used SentiSum to analyze thousands of support tickets per month. They were able to attribute conversations to respective agents and used this information to coach support team members on how to:

  • Handle difficult conversations
  • Deal with negative feedback
  • Provide extra resources for the customer

Use Case #4: Improve products based on customer feedback

Product teams often lack access to crucial user feedback, as they're disconnected from daily customer interactions handled by support teams.

When product teams ask for specific insights, support leaders often struggle to provide detailed answers as they don't have the hard data to rely on.

Speech analytics tools bridge this gap.

With a tool that analyzes not just phone calls, but all support conversations, support leaders can give data-driven answers about:

  • Features customers love or hate
  • Areas of improvement
  • New feature requests 

🎙️Related podcast: How customer support can help create better products 

SentiSum recently worked with the B2B support team at Hotjar to centralize their customer feedback and make it self-serviceable across departments. 

They now use customer insights to inform their conversations with product teams, and draw up improvement plans based on what customers are saying. 

Hotjar also plans to give their product team their own login to the SentiSum dashboard, so they can find detailed insights into product issues and improvement opportunities.

Within just a few weeks of implementing SentiSum, Hotjar’s Director of Support Nick Moreton became an evangelist of our platform. 

Nick Moreton, Director of Support @ Hotjar

Use Case #5: Get buy-in from other departments

Customer support leaders often have a “hunch” about the top issues customers are facing, customer sentiment, and agent performance.

But without the hard data to back up your claims, it can be challenging to push for improvements, especially when other teams are involved.

With a speech analytics tool like SentiSum, you can build evidence-based cases and instantly answer ad-hoc questions that other teams might have.

For instance:

  • What are the key issues customers are contacting us for? [You can answer with: 83% of our customers are contacting us about X, X, and X.]
  • What are the most important issues we need to fix first?
  • What are our customers most happy about? Who should we reach out for reviews?

Before SentiSum, Scandinavian Biolabs were receiving thousands of tickets per month and spending a lot of time manually trying to understand which of these tickets were the top issues amidst all the noise.

SentiSum's analysis revealed that support inquiries about product pricing repeatedly mentioned the lack of self-serve documentation. 

After addressing these findings with relevant teams and enhancing documentation, generic pricing queries dropped by 19%.

Anders Reckendorff, CEO and Co-Founder @ Scandinavian Biolabs

What to Look For in Speech Analytics Software

Not all speech analytics platforms are created equal, however.

At SentiSum, we believe that a good speech analytics platform should make granular insights easy to access and understand.

That comes down to four key factors:

Factor 1: Does the tool analyze data from all customer feedback channels?

While speech analytics is all about extracting insights from voice calls, a great analytics tool will bring together data from different channels, like: 

  • support tickets
  • email threads
  • phone conversations
  • social media posts
  • customer feedback surveys
  • product review platforms

Factor 2: Are the insights genuinely useful and trustable?

Great speech analytics tools turn large volumes of raw conversations into helpful, quantifiable insights. 

Actionable insights are accurate, granular, and easy to access. 

We recommend heading into your purchase decision with a few questions you’d like answered about your customers. If the tools can give you those answers confidently, that’s a great sign.

Factor 3: Is the tool easy to use by anyone who needs it?

The best speech analytics analytics software is simple and easy to use. Anyone (not just the CS team) should be able to just log in and start reading insights right away. 

This is critical to make sure the insights you find get implemented organization-wide.

Factor 4: Does the tool integrate with your existing tech stack?

Your voice of customer data likely comes from tools like Zendesk, SurveyMonkey, and Reviews (or the equivalent competitors). 

We recommend making sure that these integrations are seamless and two-way. You should be able to not just get data from these platforms, but also push automations to them.

3 Best Speech Analytics Software For Call Centers

If you're looking for a speech analytics tool for your call center, here are the 3 best ones in the market as of 2024.

Full disclosure: our speech analytics analytics software, SentiSum is one of them.

1. SentiSum: Best for Improving Customer Experience

SentiSum is an AI-powered customer analytics platform that provides detailed insights on support conversations from multiple sources—including voice calls, emails, chats, and surveys. 

How it Works

SentiSum uses machine learning-based AI to understand and analyze qualitative text from any voice of the customer channel.

In a nutshell, here’s how it works:

  • Natural language processing (NLP) technology consumes and analyzes your voice of customer data.
  • Machine learning applies granular tags on customer sentiment, topics, keywords, and more.
  • A simple yet customizable dashboard centralizes insights from all channels for easy access from Customer Support leaders and managers.
  • ChatGPT-like interface allows you to easily dig into trends and patterns in customer interactions

Who it’s For

SentiSum is best suited for medium-to-large D2C businesses and B2B tech startups with at least 3,000 support conversations per month.

Some of our best clients are customer experience leaders, product managers, and marketing or sales executives at companies like Unilever, Gousto, Ticketmaster, and Hotjar.

2. Calabrio: Best for Agent Training

Calabrio is a comprehensive platform that centralizes all customer experience evaluation tools to help improve agent performance. It serves as a command center, providing access to audio and screen recordings, evaluation forms, and key performance indicators in a unified view.

How it Works

Calabrio captures customer interactions across multiple channels, allowing users to access and review audio and screen recordings alongside evaluation forms. With its easy-to-use UI, agents, supervisors, and managers can easily navigate and access the tools they need.

Key features include:

  • 100% call recording with pause and resume for compliance,
  • Screen capture for a holistic view of each interaction,
  • Live monitoring of agent audio and desktop activity,
  • Customizable evaluation forms along with KPIs,
  • Speech analytics integration to optimize evaluations.

Calabrio also offers a suite of other tools under the umbrella of workforce performance management. For example, there are separate modules for call recording, bot analytics, and call center management.

Who it’s For

Calabrio’s customer base is made of support teams across leading brands in financial services, education, healthcare, travel, and more. It’s been used by CX teams at Netflix, Patagonia, Webhelp, and others. 

3. NICE CXone: Best for Contact Center Management

NICE CXone is a cloud-based customer experience platform that unifies omnichannel routing, analytics, workforce optimization, automation, and AI. A key component is its advanced speech analytics, which uses AI to analyze 100% of voice interactions to uncover insights for contact centers.

How it Works

NICE's speech analytics uses advanced AI models to analyze voice interactions while keeping compliance and security at the center. Here's how it works:

  • Voice interactions are captured and processed in real-time or batch mode.
  • Automatic speech recognition converts the audio into text transcripts.
  • NLP and machine learning (ML) models analyze the transcripts to uncover topics and sentiments.
  • Real-time speech analytics provides immediate feedback and guidance to agents during live calls.
  • Dashboards visualize the insights for supervisors and managers.
  • Reports are used to drive agent coaching, process improvements, and product enhancements.
  • Speech analytics data is integrated with interaction analytics for omnichannel insights.

Apart from speech analytics, NICE offers a suite of tools to manage call center operations under its flagship product, NICE CXone. Some other key features include self-service customer care, workforce engagement, and real-time agent assist.

Who it’s For

NICE CXone is a favorite among CX teams at enterprise-level organizations with 10,000+ employees and $1 billion in revenue. Some of its clients include Infosys, Cognizant, and Arizona State University.

Closing Thoughts

This article explained speech analytics: what it is, how it works, and how it helps improve customer service and agent performance. We also looked at important features to look for in these tools and highlighted three leading tools in the market.

In short, the best speech analytics software should:

  • Integrate with all your support channels
  • Use machine-learning NLP to analyze speech data
  • Come with an AI tool that answers questions about your data
  • Let you build custom AI models and automations

SentiSum ticks all these boxes, offering a comprehensive solution for speech analytics. Book a demo with us below!

SentiSum's CTA
Join a community of 2139+ customer-focused professionals and receive bi-weekly articles, podcasts, webinars, and more!

Trending articles

Customer Experience

Speech Analytics for Call Centers: 5 Use Cases & Tools

March 29, 2024
Ben Goodey
Customer Service Researcher
In this article
Understand your customer’s problems and get actionable insights
See pricing

Is your AI accurate, or am I getting sold snake oil?

The accuracy of every NLP software depends on the context. Some industries and organisations have very complex issues, some are easier to understand.

Our technology surfaces more granular insights and is very accurate compared to (1) customer service agents, (2) built-in keyword tagging tools, (3) other providers who use more generic AI models or ask you to build a taxonomy yourself.

We build you a customised taxonomy and maintain it continuously with the help of our dedicated data scientists. That means the accuracy of your tags are not dependent on the work you put in.

Either way, we recommend you start a free trial. Included in the trial is historical analysis of your data—more than enough for you to prove it works.

Call centers get thousands of phone calls a day, but only 2% of them turn into useful insights.

The other 98%? They go into a black box. 

Historically, it’s so difficult to extract insights from speech because of the sheer volume and the qualitative nature of it.

Call center leaders would either:

  1. Manually listen to a sample of the calls and tag each with keywords. This is time-consuming and not comprehensive enough. You also don’t have confidence in the evidence to send it to other teams to push for changes. 
  1. Do a huge diligent analysis once a year. This takes a technical skill set to do AND to understand the output. By the time you get insights, they’re already obsolete. 

So, what’s the solution?

Enter: Speech analytics.

In this article, we’ll go into what speech analytics is, how it works, and how you can use these insights to improve customer experience and call center performance.

Jump to:

TL;DR

1. Speech analytics for call centers is the process of analyzing recorded conversations between agents and customers using AI to extract useful insights.

2. Call centers use speech analytics to identify areas of friction in the customer journey, improve CSAT scores, assess agent performance and ensure high call quality, optimize the product roadmap, and get buy-in from other departments.

3. When looking for a speech analytics tool, prioritize one that analyzes all feedback channels, offers granular insights, is easy to use, and comes with integration support.

4. The three best speech analytics software in 2024 are SentiSum, Calabrio, and NICE CXone.

What is Speech Analytics for Call Centers?

Speech analytics for call centers is the process of analyzing support calls at scale using AI. 

AI models like Natural Language Processing listens to each call (much like a human would) and labels that conversation with a predetermined tagging taxonomy.

This includes:

  • The main topic of the call
  • Customer sentiment (e.g. Happy, sad, angry)
  • Priority (e.g. Urgent, not urgent)

For instance, let’s say 100 customers contacted you about damaged packaging during the delivery process. AI listens to, transcribes, and categorizes these calls and gives you quantifiable insights, such as:

  1. Topic: Damaged Packaging During Delivery
  2. Sentiment: 94% Negative
  3. Priority: Urgent

On a speech analytics platform like SentiSum, you can also see whether these issues are increasing or decreasing over time. If, for example, 110 customers are now contacting you about the same issue, you can quickly take action before it becomes 200 or 300. 

Source: SentiSum’s Speech Analytics Dashboard

Here are some other ways a speech analytics platform helps call center leaders:

  1. Understand agent performance at a glance. A speech analytics tool transcribes each individual support call so you can understand how each agent handled a particular situation. This information is crucial to build agent training modules and improve call quality as a whole. 
  2. Analyze support conversations from all channels. Most speech analytics tools bring in data from all your support channels, such as emails, chats, CSAT/NPS surveys, reviews, and more. This helps you understand the main issues customers are contacting you across all your channels and the sentiment behind them.
  3. Uncover the main drivers behind CSAT scores. Speech analytics software helps correlate CSAT scores with past interactions on all channels. This gives more context to low or high scores.
  4. Build evidence-based cases for change. Speech analytics platforms turn qualitative calls into quantitative data at scale. This data is crucial for call center leaders to push for improvements in other departments and answer ad-hoc questions they may have.
Companies that leverage speech analytics saved 20% to 30% in cost, improved CSAT scores by 10% or more, and saw stronger sales. Source: McKinsey & Company

So, how does a speech analytics tool work exactly?

How Call Center Voice Analytics Works: The Tech Behind

Call center voice analytics tools use AI-powered natural language processing (NLP) to analyze large volumes of conversations between agents and customers. 

In a nutshell, here’s how it works:

  • First, the software records each support call.
  • It then uses speech recognition technology to transcribe the spoken words into text.
  • A NLP model then analyzes the transcribed text, uncovering key topics and sentiment.

To give you a full picture of how speech analytics software works, we’ll use our own tool–SentiSum– to illustrate the process. 

(For the visual learners, here’s a 30-second YouTube video of a summary breakdown. We recommend watching it at 0.5x speed. ⬇)

Part 1. Integration with your existing help desk platforms. 

The speech analytics software first brings in voice call data from your helpdesk platforms. For example, SentiSum integrates with most major (and minor) helpdesk platforms like:

  • Zendesk
  • Freshdesk
  • Gorgias 

You can explore the full integration list here

But apart from voice calls, SentiSum also brings in data from all other feedback channels, such as surveys, customer support chats, emails, and more, onto one dashboard.

Part 2. AI analyzes voice calls. 

Instead of listening to calls manually yourself or having your agents add tags, SentiSum automatically analyzes and applies detailed tags using machine learning-based artificial intelligence.

In the example below, you can see how powerful machine learning NLP is compared to keyword extraction (what platforms like Zendesk typically use) and rule-based NLP (what other software providers use).

Learn more about different types of NLPs here.

Different types of NLP models - SentiSum
Different types of NLP models

Part 3. See customer call topics and trends. 

Once the analysis and tagging process is done (usually takes seconds), SentiSum’s easy to use dashboard shows all the key insights, like:

  • Trend of voice call volume over time 
  • Main issues for contact (we call them “topics”)
  • Top increases and decreases in sentiment and volume
A screenshot of SentiSum's insights dashboard - SentiSum
SentiSum's insights dashboard

If you click into each topic (e.g. “Request Refund”), you can see:

  • All conversations tagged under the topic
  • Increases and decreases over time
  • AI chatbot called “Dig In” - which allows you to ask questions and get succinct, summarized answers.
A screenshot of SentiSum's dashboard - SentiSum
Detailed breakdown each topic

Want to dig into how each support call went? No problem.

You can click into each call conversation and get:

  • An AI summary of each call 
  • Agent performance analysis
  • Transcript of the call

Part 5. Insights are pushed to your help desk.

These insights shouldn’t just sit in your analysis tool, but also be pushed back to your customer facing platform. 

On SentiSum, you can create automation workflows using a custom AI model and push triage and prioritization rules back to your helpdesk platform. 

This helps your agents understand which issues to fix first based on sentiment and urgency (e.g. angry customers need to be attended to quickly).

Want to see what the platform looks like and how it can help you? Please book a demo below!

SentiSum's CTA

How Do Call Centers Use Speech Analytics? 5 Use Cases to Improve Customer Experience and Agent Performance

So, you’ve got the insights. Now what? 

Here are five ways you can use the insights from speech analytics to improve your product, customer experience, and teams. 

Use Case #1: Identify Friction in the Customer Journey

A customer calling your service team represents friction. Whether it’s a delivery gone wrong or a damaged product, these are issues customers can’t fix themselves.

These issues, if they become a bad-enough experience, can cause customers to churn and choose competitors in the future, leading to revenue loss.

The key is to identify these friction points early on before they escalate. This way you can draw up plans to fix the issues and improve customer experience, driving loyalty and revenue in the future.

On SentiSum, you can see the main topics that are decreasing or increasing in volume over time (daily, weekly, or monthly).

This helps you quickly understand (and quantify) which issues you should prioritize fixing and use this data to push for improvements in other departments. 

For example, Lakrids, a gourmet D2C company, used SentiSum to understand what customers are saying across their Dixa support tickets. 

They noticed that towards the end of 2021, there was an influx of tickets regarding damaged packaging during delivery.

Based on this insight, their Head of Customer Service and Sales Support drew up a plan to improve packaging and transportation, along with a compensation plan for affected customers.

This initiative led to a 26% decrease in packaging-related complaints within a year.

Løkke Engraf, Head of Consumer Service and Sales Support @ Lakrids

Use Case #2: Improve your CSAT scores

Call centers usually send out CSAT surveys at the end of each support conversation.

But CSAT scores mean very little without context. Without being able to correlate CSAT data with your support tickets and agent conversations, you have to rely on hunches to figure out what went wrong.

SentiSum maps your CSAT scores to your voice call topics automatically, so you understand why a customer gave you a low (or high) score even if they didn’t leave a comment. This is key to fixing issues at the very root and improving CSAT scores in the future.

You can also use these explanations to coach your agents to provide better customer service. 

🎙️Related podcast: How to drive customer service CSAT from 30% to 90% 

Going back to Lakrids, our platform helped make sense of the vast amount of disparate customer data they had in their support platform, Dixa. 

By connecting each CSAT score with matching support tickets, we helped them understand why customers left the feedback they did. 

In less than 12 months, Lakrids improved their CSAT scores by 9%. 

Løkke Engraf, Head of Consumer Service and Sales Support @ Lakrids

Use Case #3: Ensure agent quality.

Training agents is hard when you don’t have the full picture of what’s going on.

If you’re randomly analyzing a small section of agent calls, you’re likely to miss out on broader trends of agent performance. 

Speech analytics can help you implement tighter standards for quality assurance by analyzing ALL your support calls.

For instance, SentiSum shows you the transcripts of each call under each topic. You can click into each to understand what went wrong and what went well. 

Our ChatGPT-like AI also gives you a summary of how each call went, so you don’t need to read the entire conversation. We also recently launched ‘Analysis’ (Beta) which gives you a breakdown of agent performance. 

Together, they provide detailed feedback on how specific agents handle each interaction. For example:

  • Was the issue resolved? How long did it take to be resolved?
  • Was the agent polite and empathetic towards the customer?
  • Was the agent professional in handling the call? What could they have done better?

You can then use these insights to set up agent training modules or to improve a particular agent’s performance. 

For instance, Scandinavian Biolabs, a haircare D2C brand, used SentiSum to analyze thousands of support tickets per month. They were able to attribute conversations to respective agents and used this information to coach support team members on how to:

  • Handle difficult conversations
  • Deal with negative feedback
  • Provide extra resources for the customer

Use Case #4: Improve products based on customer feedback

Product teams often lack access to crucial user feedback, as they're disconnected from daily customer interactions handled by support teams.

When product teams ask for specific insights, support leaders often struggle to provide detailed answers as they don't have the hard data to rely on.

Speech analytics tools bridge this gap.

With a tool that analyzes not just phone calls, but all support conversations, support leaders can give data-driven answers about:

  • Features customers love or hate
  • Areas of improvement
  • New feature requests 

🎙️Related podcast: How customer support can help create better products 

SentiSum recently worked with the B2B support team at Hotjar to centralize their customer feedback and make it self-serviceable across departments. 

They now use customer insights to inform their conversations with product teams, and draw up improvement plans based on what customers are saying. 

Hotjar also plans to give their product team their own login to the SentiSum dashboard, so they can find detailed insights into product issues and improvement opportunities.

Within just a few weeks of implementing SentiSum, Hotjar’s Director of Support Nick Moreton became an evangelist of our platform. 

Nick Moreton, Director of Support @ Hotjar

Use Case #5: Get buy-in from other departments

Customer support leaders often have a “hunch” about the top issues customers are facing, customer sentiment, and agent performance.

But without the hard data to back up your claims, it can be challenging to push for improvements, especially when other teams are involved.

With a speech analytics tool like SentiSum, you can build evidence-based cases and instantly answer ad-hoc questions that other teams might have.

For instance:

  • What are the key issues customers are contacting us for? [You can answer with: 83% of our customers are contacting us about X, X, and X.]
  • What are the most important issues we need to fix first?
  • What are our customers most happy about? Who should we reach out for reviews?

Before SentiSum, Scandinavian Biolabs were receiving thousands of tickets per month and spending a lot of time manually trying to understand which of these tickets were the top issues amidst all the noise.

SentiSum's analysis revealed that support inquiries about product pricing repeatedly mentioned the lack of self-serve documentation. 

After addressing these findings with relevant teams and enhancing documentation, generic pricing queries dropped by 19%.

Anders Reckendorff, CEO and Co-Founder @ Scandinavian Biolabs

What to Look For in Speech Analytics Software

Not all speech analytics platforms are created equal, however.

At SentiSum, we believe that a good speech analytics platform should make granular insights easy to access and understand.

That comes down to four key factors:

Factor 1: Does the tool analyze data from all customer feedback channels?

While speech analytics is all about extracting insights from voice calls, a great analytics tool will bring together data from different channels, like: 

  • support tickets
  • email threads
  • phone conversations
  • social media posts
  • customer feedback surveys
  • product review platforms

Factor 2: Are the insights genuinely useful and trustable?

Great speech analytics tools turn large volumes of raw conversations into helpful, quantifiable insights. 

Actionable insights are accurate, granular, and easy to access. 

We recommend heading into your purchase decision with a few questions you’d like answered about your customers. If the tools can give you those answers confidently, that’s a great sign.

Factor 3: Is the tool easy to use by anyone who needs it?

The best speech analytics analytics software is simple and easy to use. Anyone (not just the CS team) should be able to just log in and start reading insights right away. 

This is critical to make sure the insights you find get implemented organization-wide.

Factor 4: Does the tool integrate with your existing tech stack?

Your voice of customer data likely comes from tools like Zendesk, SurveyMonkey, and Reviews (or the equivalent competitors). 

We recommend making sure that these integrations are seamless and two-way. You should be able to not just get data from these platforms, but also push automations to them.

3 Best Speech Analytics Software For Call Centers

If you're looking for a speech analytics tool for your call center, here are the 3 best ones in the market as of 2024.

Full disclosure: our speech analytics analytics software, SentiSum is one of them.

1. SentiSum: Best for Improving Customer Experience

SentiSum is an AI-powered customer analytics platform that provides detailed insights on support conversations from multiple sources—including voice calls, emails, chats, and surveys. 

How it Works

SentiSum uses machine learning-based AI to understand and analyze qualitative text from any voice of the customer channel.

In a nutshell, here’s how it works:

  • Natural language processing (NLP) technology consumes and analyzes your voice of customer data.
  • Machine learning applies granular tags on customer sentiment, topics, keywords, and more.
  • A simple yet customizable dashboard centralizes insights from all channels for easy access from Customer Support leaders and managers.
  • ChatGPT-like interface allows you to easily dig into trends and patterns in customer interactions

Who it’s For

SentiSum is best suited for medium-to-large D2C businesses and B2B tech startups with at least 3,000 support conversations per month.

Some of our best clients are customer experience leaders, product managers, and marketing or sales executives at companies like Unilever, Gousto, Ticketmaster, and Hotjar.

2. Calabrio: Best for Agent Training

Calabrio is a comprehensive platform that centralizes all customer experience evaluation tools to help improve agent performance. It serves as a command center, providing access to audio and screen recordings, evaluation forms, and key performance indicators in a unified view.

How it Works

Calabrio captures customer interactions across multiple channels, allowing users to access and review audio and screen recordings alongside evaluation forms. With its easy-to-use UI, agents, supervisors, and managers can easily navigate and access the tools they need.

Key features include:

  • 100% call recording with pause and resume for compliance,
  • Screen capture for a holistic view of each interaction,
  • Live monitoring of agent audio and desktop activity,
  • Customizable evaluation forms along with KPIs,
  • Speech analytics integration to optimize evaluations.

Calabrio also offers a suite of other tools under the umbrella of workforce performance management. For example, there are separate modules for call recording, bot analytics, and call center management.

Who it’s For

Calabrio’s customer base is made of support teams across leading brands in financial services, education, healthcare, travel, and more. It’s been used by CX teams at Netflix, Patagonia, Webhelp, and others. 

3. NICE CXone: Best for Contact Center Management

NICE CXone is a cloud-based customer experience platform that unifies omnichannel routing, analytics, workforce optimization, automation, and AI. A key component is its advanced speech analytics, which uses AI to analyze 100% of voice interactions to uncover insights for contact centers.

How it Works

NICE's speech analytics uses advanced AI models to analyze voice interactions while keeping compliance and security at the center. Here's how it works:

  • Voice interactions are captured and processed in real-time or batch mode.
  • Automatic speech recognition converts the audio into text transcripts.
  • NLP and machine learning (ML) models analyze the transcripts to uncover topics and sentiments.
  • Real-time speech analytics provides immediate feedback and guidance to agents during live calls.
  • Dashboards visualize the insights for supervisors and managers.
  • Reports are used to drive agent coaching, process improvements, and product enhancements.
  • Speech analytics data is integrated with interaction analytics for omnichannel insights.

Apart from speech analytics, NICE offers a suite of tools to manage call center operations under its flagship product, NICE CXone. Some other key features include self-service customer care, workforce engagement, and real-time agent assist.

Who it’s For

NICE CXone is a favorite among CX teams at enterprise-level organizations with 10,000+ employees and $1 billion in revenue. Some of its clients include Infosys, Cognizant, and Arizona State University.

Closing Thoughts

This article explained speech analytics: what it is, how it works, and how it helps improve customer service and agent performance. We also looked at important features to look for in these tools and highlighted three leading tools in the market.

In short, the best speech analytics software should:

  • Integrate with all your support channels
  • Use machine-learning NLP to analyze speech data
  • Come with an AI tool that answers questions about your data
  • Let you build custom AI models and automations

SentiSum ticks all these boxes, offering a comprehensive solution for speech analytics. Book a demo with us below!

SentiSum's CTA

Frequently asked questions

Is your AI accurate, or am I getting sold snake oil?

The accuracy of every NLP software depends on the context. Some industries and organisations have very complex issues, some are easier to understand.

Our technology surfaces more granular insights and is very accurate compared to (1) customer service agents, (2) built-in keyword tagging tools, (3) other providers who use more generic AI models or ask you to build a taxonomy yourself.

We build you a customised taxonomy and maintain it continuously with the help of our dedicated data scientists. That means the accuracy of your tags are not dependent on the work you put in.

Either way, we recommend you start a free trial. Included in the trial is historical analysis of your data—more than enough for you to prove it works.

Do you integrate with my systems? How long is that going to take?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

What size company do you usually work with? Is this valuable for me?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

What is your term of the contract?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

How do you keep my data private?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Support Insights Community
Join a community of 2200+ customer-focused professionals and receive bi-weekly articles, podcasts, webinars, and more!

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Speech Analytics Call Center FAQs

What is the difference between voice and speech analytics?

While the two terms are often used interchangeably, there are subtle differences between voice and speech analytics.

Speech analytics focuses primarily on the words spoken during a voice interaction. It involves transcribing audio, analyzing the choice of words, and identifying topics. Voice analytics, on the other hand, analyzes how the words were spoken by examining audio patterns like tone, pitch, stress, tempo, and rhythm. 

How to implement speech analytics in a call center?

Here’s a 5-step process for quickly implementing speech analytics in a call center:

  • Identify the key areas you want to improve, such as customer satisfaction, agent performance, or operational efficiency.
  • Select a solution that aligns with your requirements, provides advanced capabilities like sentiment analysis and categorization, and integrates well with your existing call center software.
  • Ensure you have a comprehensive dataset of recorded calls with accurate timestamps. 
  • Teach the system to recognize various speech patterns, languages, and accents used by your agents and customers. Adjust settings wherever necessary.
  • Regularly review the generated insights to identify recurring patterns, areas for improvement, and necessary actions. Use the data to create reports, provide feedback to agents, and implement process changes.

What’s the future of speech analytics?

Advances in AI and machine learning (ML) have been a key driver of innovation for speech analytics in the last few years. With more accurate NLP models that can understand human language, analytics tools have gotten better at understanding sentiments or topics, predicting trends through historical data, and offering real-time recommendations for agent performance.

Is your AI accurate, or am I getting sold snake oil?

The accuracy of every NLP software depends on the context. Some industries and organisations have very complex issues, some are easier to understand.

Our technology surfaces more granular insights and is very accurate compared to (1) customer service agents, (2) built-in keyword tagging tools, (3) other providers who use more generic AI models or ask you to build a taxonomy yourself.

We build you a customised taxonomy and maintain it continuously with the help of our dedicated data scientists. That means the accuracy of your tags are not dependent on the work you put in.

Either way, we recommend you start a free trial. Included in the trial is historical analysis of your data—more than enough for you to prove it works.