Speech analytics

There's so much valuable insight buried in each customer interaction, but the question is how to extract that information. That's where speech analytics comes in.

For most companies, getting useful business intelligence and insights that help them understand their customers better doesn’t require fancy surveys or outsourcing work to expensive research agencies.

They already have all the insights they need—in their customer calls. And their contact center teams probably already have easy access to this information.

Traditionally, it would be through listening to call recordings one by one and taking notes, which, as you can imagine, is pretty time-consuming and labor-intensive.

Good news is, call centers and contact centers don’t have to do that anymore because of automation. Specifically, speech analytics.

But what exactly is speech analytics? How does it work? And how can it help contact centers improve operational efficiency?

What is “speech analytics?”

As its name suggests, speech analytics are the data and insights you get from digging into conversations that your business has—usually on the phone, and usually with customers and prospects.

You can get speech analytics for both live calls and recorded ones, and if your software has a really good speech-to-text feature, it could even transcribe calls in real-time.

For example, Dialpad’s Vi (Voice Intelligence) technology does exactly this—but can also analyze speech and pick up on keywords and phrases chosen by the user.

This functionality gives contact centers many more possibilities in terms of what they can do with all these conversations they’re having with customers every day.

It also integrates with CRMs like Salesforce and HubSpot, and other tools like Zendesk to provide these real-time transcriptions and automatically log customer interactions, right inside those tools:


How do speech analytics work?

Think of all the data your business accumulates each day through phone calls and emails. That data is basically unstructured and not always useful in its current form.

For example, having just the transcript or recording of a customer call is not very useful. You also need the metadata associated with that speech, like who said it, when it was said, and so on. That’s what you’d need speech analytics software for.

Typically, it would put the audio through an automatic speech recognition process to transcribe the conversation from speech to text. From there, it can analyze what was said to tell you, for instance, whether the sentiment on that call was negative or positive.

Dialpad’s Contact Center can do this—in real time, as the call is happening:

👉Dialpad tip: A good speech analytics solution should at the very least be able to transcribe conversations accurately. Otherwise, it can’t provide you with the actionable insights you need to make informed decisions and improve agent performance.


What can contact centers do with speech analytics?

Spot “problem areas” more easily

Are certain competitors coming up in conversations more? Are there new features or products that customers are asking about? Maybe you’re interested in reasons why customers call to ask for refunds. You can program these words and phrases into your speech analytics or contact center platform so that it’ll track how often they come up in conversations and provide a better customer experience.

In Dialpad, you can quickly search call transcriptions for keywords important to you and your business:


Reduce churn

Churn rates can vary by industry, but often sit around 20-25%. That represents a lot of lost customers and revenue. The challenge with understanding your churn is that the reasons why customers leave can vary greatly, from poor customer service to dissatisfaction with the actual products.

While a speech analytics tool may not solve every issue behind customer churn, it can go a long way in helping you understand what these issues are.

For example, in Dialpad you can create Custom Moments to track how often specific keywords and phrases come up in calls:


All you have to do is create a Custom Moment—let’s say to help you track subscription cancellations—and set it to trigger whenever customers say “money back” or “cancel” or “refund” on phone calls:


Better reporting

Knowing what’s being said in customer conversations is good, but what if you need to report on metrics and KPIs? Dialpad, for example, can show you easy-to-understand graphs of how sentiment is trending and which keywords are recurring frequently:


Coach agents at any time

Often you’ll take a look at data after a call, but what if an agent needs assistance during a phone call? Call center managers and supervisors can listen in on, barge, and take-over phone calls to coach agents while they’re speaking to customers. Since customer satisfaction shifts in seconds, design a call center that prioritizes immediate action through live sentiment analysis and real-time assist cards.

Provide a great customer experience while lowering costs

If you can get better customer insights from your calls, then you can improve the way you handle those calls. What this means in real terms is that you can reduce the number of repeat calls, improve your quality assurance and quality monitoring efforts, and ultimately lower costs in areas such as staffing.

Easily identify and address compliance issues

There are many industries where compliance is a top priority for contact centers and call centers. Many businesses have to adhere to legislation like the Payment Card Industry Data Security Standard (PCI-DSS) and Healthcare Insurance Portability and Accountability Act (HIPAA).

As well as those specific regulations, you also need to adhere to the requirement in many states that customers must give consent for a call to be recorded. With speech analytics, you can identify any issues with compliance in your customer engagements more easily and address them through individual or group training. Dialpad also offers enterprise level security in its Vi system to ensure data protection will never be an issue.

FAQs about speech analytics

Which types of businesses can benefit from speech analytics?

While contact centers and call centers are the main beneficiary of speech analytics, any industry that has regular phone communication with customers can also benefit. (And you can combine it with other conversations on, say, email and live chat too for an omnichannel analysis.) That includes banks and financial organizations, utility companies, the travel and hospitality sector, healthcare organizations, and more. (Yes, even businesses with more strict compliance needs!)

What are the signs that a company is ready to invest in speech analytics and voice intelligence?

Every company is different of course, but there are a few signs that it’s time to consider implementing speech analytics. For example, you may be struggling with some aspects of compliance, your contact center metrics may be down, or you’re seeing increased churn—without having insights into why this is happening. You might also just want to improve quality management or agent performance in general, and need to find the gaps in your customer journey first.

How does speech analytics encourage better team training?

Contact centers often have a high turnover rate, and constantly training new staff costs time and money. Speech analytics can not only improve training when onboarding new staff, it can also help with ongoing coaching as new issues and questions come up. And if it helps them do their job better, then it could also improve employee satisfaction and reduce staff turnover.

Can speech analytics also help measure agent KPIs?

Yes, speech analytics allow you to be more data-driven—without having to listen to as many calls. You can track call sentiment on your agents’ calls, and see if any problematic keywords and phrases come up more frequently for certain agents than others.

Can I use speech analytics for marketing too?

Of course. With all the data you get from speech analytics software, you should be able to identify trends in customer interactions, behaviors, and desires that can help with optimizations to your marketing strategy. That includes brand mentions, competitor intelligence. Just as you can track the sentiments attached to any mentions of your brand, and more.

I don’t run a big contact center. Do the benefits of speech analytics apply to me?

It depends! If you have enough phone calls on a daily basis with customers and you believe there are untapped insights, then you may benefit from having speech analytics. The benefits are pretty far-reaching, from improved customer experiences to increased sales—it’s more about how creative you can be with the insights that you uncover.

What is the best speech analytics software for my business?

Dialpad’s Vi is, in our humble opinion, the ideal choice. Its unparalleled natural language processing (NLP), recognition of phonetics, and machine learning makes it more accurate than Google’s transcription—and what’s unique about Dialpad is that it can provide insights in real time—not after the call. If you want to gain a better understanding of your customer relationships, it’s a great option.

Start harnessing AI-driven speech analytics with Dialpad

You may already have a team of great agents working in your contact center or call center, but with a good speech analytics tool, you can turn those great agents into super agents.

By analyzing the conversations between customers and agents, either in real-time or later as call recordings or transcriptions, you can adjust your tactics to improve your contact center performance.