Conversational AI

Speech AI Spotlight: Reimagine Customer Service with Virtual Agents

Customer service representatives working

Virtual agents or voice-enabled assistants have been around for quite some time. But in the last decade, their usefulness and popularity have exploded with the use of AI.

According to Gartner, virtual assistants will automate up to 75% of tasks for call center agents by 2025–up from 30% in 2021. This translates to a better experience for both contact center agents and customers.

From healthcare to financial services, AI has transformed customer service making it more efficient and personalized. Today’s virtual agents, powered by speech AI technology, handle not only repetitive customer requests in contact centers, but also assist human agents in solving complex questions faster than ever before.

Based in Los Angeles, NVIDIA partner Gridspace is a voice technology and AI software company creating natural-sounding virtual agents and voice bots to enhance the customer service experience. They are also a member of the NVIDIA Inception Program, which helps startups evolve by providing access to cutting-edge technology and NVIDIA experts.

A study conducted by a Fortune 10 healthcare company reported that 72% of consumers found Gridspace virtual agents sounded more human-like than a leading and well-known company’s virtual agents.

Sign up for the latest Speech AI News from NVIDIA.

Virtual agents for complex contact centers 

To deliver great customer service over the phone, human agents must be able to talk and communicate effectively. Even the best live agents in a contact center have limited time and context when helping a caller.

Virtual agents have infinite availability, capacity for context, and consistent communication skills. They can also route calls to designated specialists.

Most of the Gridspace virtual agents operate in the financial and healthcare industries. They must be capable of handling complex tasks that require a high level of accuracy, which is extremely important when a customer’s money, health, or emotional well-being is involved.

For instance, virtual agents can assist with the patient discharge processes in hospitals. They can also assist nurses by providing prior information about a patient’s wellness and any discharge-related instructions.

In these demanding situations, the virtual agents complete calls for more than 70% of the patients they reach, which is comparable to the performance of human administrators and nurses.

Video 1: A virtual agent and human agent interact while providing customer service

Advancements in speech AI technology have enabled a new breed of virtual agents in contact centers that engage in natural conversations, understand industry jargon, and execute more tasks. 

By using a natural-sounding voice to greet callers, virtual agents can establish a strong rapport with a customer to help them feel more at ease. With cutting-edge speech models and GPU-optimized inference techniques, Gridspace virtual agents sound human-like and run in real time.

Every industry also has its own vernacular. With the latest advancements in ASR and modern training pipelines, virtual agents can understand domain-specific conversational speech and reduce resolution times to new customer problems. Additionally, virtual agents can eliminate hold times by calling customers back when they are ready. This means more patients or customers can be served day or night regardless of their location or the language they speak.

Enhancing the customer service experience

Gridspace virtual agents and voice observability software are powered by NVIDIA GPUs. Because GPUs can handle thousands of requests on a large scale, virtual agents can continue to serve customers efficiently.

The company uses GPUs, including the NVIDIA T4 and NVIDIA A100 for training and inference, serving over 50 K concurrent calls. They are also used to power text-to-speech models for conversational analysis. This includes emotion recognition, dialog actions, voice biometrics, and multimodal conversational AI.

“NVIDIA GPUs accelerate neural network computations and empower Gridspace real-time speech and language models at enterprise scale,” said Cooper Johnson, voice designer at Gridspace.

Future of virtual agents across industries

Contact centers are at the center of every global business. Virtual agents can both empower live agents and facilitate a smoother customer service experience across industries

Virtual agents reduce the likelihood of costly operational mistakes or missed opportunities by handling repetitive tasks, such as onboarding customers or making follow-up calls. They free up live agents to address more complex issues that require human-to-human interactions. 

Recently, Gridspace deployed Grace, a voice-enabled assistant that skillfully closes gaps in patient care and new customer onboarding processes. In a natural, friendly way, Grace can handle multi-turn customer interactions. 

Empower your contact center agents

Interested in adding speech AI to your virtual agent application? Following are some of the resources to get started:

Discuss (0)