Giving Chatbots a Voice to Improve Customer Experiences

Charlie Mitchell

Overcome the bruising struggles of building voicebots

Insights
Giving Chatbots a Voice to Improve Customer Experiences

Love them or loathe them, bots are fast becoming the next frontier of customer experience.

Yet, some are better than others, and one particular use case gaining momentum is voicebots.

Indeed, as McKinsey & Company states: “IVR systems are evolving from dumb menu systems into smart ‘voicebots’ capable of handling complex customer queries.”

Thankfully, “handling” involves much more than mimicking a traditional IVR that routes calls and offers callbacks.

Indeed, voicebots offer an attractive alternative to the disjointed, cumbersome software well known for frustrating customers.

Why Are Voicebots the Future of CX?

First off, consider contact centre routing. Thanks to advances in NLU and automated speech recognition, voicebots grasp customer intent from in just one or two words – without pretraining. Customers then filter into the correct queue instantaneously.

Of course, capturing intent in such a way also allows for much more accurate ticket tagging.

Yet, let’s reconsider the routing example. Sure, a voicebot can pass the customer onto agents, but it may also send them to a self-service flow powered by robotic process automation (RPA).

The result? Lightning fast service, automated customer contacts, and relieved contact centre agents.

Naturally, self-service will fail from time to time. However, as Patrice Mattez, VoiceAI Connect Business Development Manager at AudioCodes, says:

“At any time, the customer can request that the virtual assistant transfers the call to a human agent. It then does so with the transcript, metadata, conversation history and the virtual assistant.”

Mattez also suggests that when a query is too complex to automate fully, the voicebot can gather customer information – pertinent to the intent – and streamline the following customer-agent conversation.

With the right approach, companies may also design self-service flows by reapplying the logic of existing, high-satisfaction chatbots for voice.

By doing so, companies can maximize the value of their bot efforts across the omnichannel contact centre while adding more self-service flows.

In a world that is becoming accustomed to using smart speakers, customers can take advantage of these voice-driven self-service options to resolve their queries without lifting a finger.

How to Build the Best Voicebot

There are three fundamentals to building a virtual voice assistant:

  1. Conversational copywriting that stems from a deep analysis of call recordings.
  2. Psychology that aims to foster helpful service and the empathetic glow of natural conversation.
  3. Technology!

When it comes to the latter, companies often obsess over working with a single bot vendor, whether that is Microsoft, IBM, or a Gartner Magic Quadrant leader, like Kore.ai.

However, success often springs from companies remaining as open as possible when building bots. This is according to Mattez, who states:

“We have seen organizations selecting a monolithic provider and having to start everything from scratch with another provider later on as accuracy and results were very poor.”

So, where does it go wrong? In the case of many international businesses, it is in attempting to apply the same voicebot framework across multiple regions and countries.

Unfortunately, a voicebot is not as easy to design as a chatbot. In many cases, companies hire different bot frameworks for handling the same conversations.

Indeed, speech-to-text (STT) and text-to-speech (TTS) models – which form the foundation of voicebot technology – offer varying degrees of success across different accents, dialects, and languages. The accuracy of these cognitive services changes between various vendors.

Some companies have even built voicebots that switch from one provider to another during the conversation, as the second vendor offers higher accuracy for particular parts of the interaction.

Yet, for Mattez, the most significant lesson is to work with a technology partner who understands the challenges of voice and can provide open APIs to fit the organization’s requirements.

He adds that the vendor:  “Must integrate very fast with any contact centre, any bot framework, any STT and TTs and with new cognitive services such as voice authentication while language detection is also key.”

Where Does AudioCodes Come In?

AudioCodes VoiceAI Connect addresses the challenges above head-on, offering a communication hub between various voicebot frameworks – including Microsoft, Google, and AWS.

VoiceAI Connect addresses the challenges above head on, offering a communication hub between various voicebot frameworks – including Microsoft, Google, and AWS.

As a result, users can combine the best cognitive services of each framework, enabling a simpler approach to implementing voice-based virtual agents and conversational IVR. These work as an extension of a company’s existing conversational AI and telephony systems.

Harnessing the solution, companies can transform chatbots into voicebots and “let their bots do the talking.”

Making this point, Mattez states:

“It is easy to deploy our VoiceAI Connect solution in the cloud, private or public, shared or dedicated, or even in the customer Data Center. We bring flexibility and peace of mind to organizations as our solution is a fully managed service.”

The solution connects the bot with the contact centre, UC system and SIP trunk providers in minutes, enabling the flexibility that Mattez lauds.

It also offers the platform for Proof of Concepts, in addition to the production of self-service portals – while working seamlessly with Microsoft Teams and Dynamics 365-driven operations.

To discover more and kickstart your voicebot journey, visit: voiceaiconnect.audiocodes.com

 


Join our Weekly Newsletter