NLP and NLU – What’s the Difference?

NLP and NLU – What’s the Difference? (IVR Technologies Defined!)

A modern conversational IVR uses two important components to ‘listen’ to callers.

Natural Language Processing (NLP) and Natural Language Understanding (NLU) have distinct roles – but the terms are often used interchangeably.

These IVR technologies are vital to the next generation of automated call tools. So we’d better get a good idea of what they are!

In this post:

IVR technologies defined: What is NLP?

NLP is an area of Artificial Intelligence focused on turning speech into structured data. The aim is to turn human language – which is disorderly and poorly defined – into forms of data that a machine can easily process.

Here’s another way to look at the purpose of NLP: it aims to make machine intelligence more human-like. That can be a huge boon for the inbound call center

Speech is the fastest and simplest ‘output method’ for you and me. But there are a lot of tricky variations in speech for your IVR system to cope with, including:

  • Accent
  • Sentence structure
  • Word choice
  • Disfluency
  • Homophones, synonyms, metonymy and more!

These are all either variations in how humans use language, or eccentricities in how language works. Put it this way – whenever someone designs a new language from scratch, it never looks like the messy languages humans have developed over time!

IVR technologies defined: What is NLU?

NLU is an important subset of NLP. It looks beyond which words were used and determines what those words mean – what is the intent of the speaker?

This is a core part of conversational IVR; it enables the system to respond to words and phrases with appropriate actions.

Let’s say you call a speech-enabled IVR and say, ‘I want to check my reservation for June 23’. The system’s NLP understands each word, but it takes NLU to determine that ‘reservation’ is the key concept.

There are plenty of actions related to ‘reservation’ including make, check, update, cancel and so on. But you used the word ‘check’ so the system has a strong idea of what you’re trying to achieve.

Based on that it can perform the next action: telling you the details of your booking. 

(It could also perform other CSat boosting actions, like sending an automated SMS with the right information. Want to know how? Check out ‘SMS is the Unsung Hero of Call Center Technology!’) 

Is conversational IVR worth the effort?

Plenty of large contact centers already use conversational IVR. Many more will soon. But adding new features to your IVR systems costs money – so is it a good idea?

You can figure that out by asking yourself these four questions about your current service.

Do we need more customers to use self-service?

IVR systems have traditionally been a pretty weak form of self-service. DTMF inputs (push 1 for sales, etc.) are long-winded. They also make users listen to more irrelevant options than relevant options.

As a result, many contact centers use their IVR for call routing rather than self-service. And that’s a waste!

It’s already clear that customers want to use conversational tools; almost three quarters say their self-service experience is improved by devices that can talk.  

If you need more customers to use self-service, there are a lot of ways to pursue that. But the customers who have already called your contact center are some of the hardest to convert. 

Customers will use better self-service tools if you provide them. They don’t always need or want an agent.

(Interested in bringing more self-service into your contact center? Read ‘Should you replace contact center services with self-service?’) 

IVr technologies

Does our current IVR have a negative impact on CX?

There’s a good chance that your current IVR harms your customer experience. Not all IVRs are set up poorly. But IVRs tend to rank very high on lists of the things customers dislike. (They often top the list…)

customer attitudes IVR


There are three strong indicators that customers dislike your IVR:

  • They tell you! Either in surveys or by complaining to agents
  • They zero out or abandon the call entirely
  • The IVR has a poor FCR rate

It’s totally possible to fix problems in your IVR. If users regularly abandon or zero out in specific pathways, it generally indicates a cul-de-sac or similar problem.

(You can read up on IVR testing here.) 

But fixing these issues takes time, and you may just be fighting against the tide. Research shows that customers expect to use speech interfaces in around 70% of self-service interactions by 2023! 

Do we need to preserve our brand image?

One way or another, most businesses market their high-level customer service standards. There’s a good reason for that. Almost everyone – 96% of customers – say that customer service plays a key role in the choice of (and loyalty to) brands.  

So it’s sensible to talk up your service standards.

But talking the talk is only step one. You have to walk the walk as well!

If you’re offering customers a dated and hard-to-use DTMF system, that quickly undercuts the image you’re trying to present.

And this problem is only going to get worse. Gartner predicts that, as soon as 2025, around a third of businesses will use a conversation platform in customer service.  

Before too long, self-service systems that don’t use conversational IVR technologies will seem hopelessly old fashioned. They’ll be brand liabilities.

conversational IVR


Will IVR technologies cut our costs?

We started by talking about cost – conversational IVR comes with a price tag.

The question is, will the system pay for itself? That’s an open question when it comes to things like brand image and customer experience. Only you know the cash value of brand loyalty to your business.

But when it comes to things like reducing agent-handled calls and increasing overall automation, the cash savings from conversational IVR are obvious.

A surprising number of enterprise-scale businesses have directly saved millions of dollars by reducing strain on their contact centers. 

Businesses like Paypal, Wells Fargo, FedEx and Delta have already seen drastic changes to their self-service uptake. As a result, they’ve made savings ranging from $3-5 million annually.  

So, you’ve asked yourself four important questions: 

  • Do we need more customers to use self-service?
  • Does our current IVR have a negative impact on CX?
  • Do we need to preserve our brand image?
  • Do we need to reduce the cost of our contact center?

If you answered yes to even one of these questions, conversational IVR technologies may be a good fit.

If not – if you already run the perfect business – customers are going to make that decision for you in the next few years. As we’ve already seen, expectations for easy-to-use tools are growing every day. 

On top of that, virtual home assistants like Alexa are teaching a generation of consumers how to interact with machines via voice. Your customers are talking to Siri far more than they’re talking to customer service agents. 

In other words, this is the future that’s just around the corner. Better get ready for it!

Related Posts