FPT Smart Cloud Levels Up Customer Service Operations With NVIDIA
In this rapidly evolving digital era, virtual assistants are being integrated into contact centers to assist customers with both simple and sophisticated tasks. Customers expect these virtual assistants to speak as naturally and intuitively as real people. In turn, businesses are demanding a variety of AI voices they can deploy in their digital experiences.
Unleashing Next-Generation Customer Service With Artificial Intelligence
As a leading AI provider in Vietnam, FPT Smart Cloud has developed FPT AI Engage, a virtual assistant for the call center that automatically makes outbound calls, handles inbound calls, and directs calls with smart interactive voice response (IVR).
FPT AI Engage is grounded on four core technologies: large language models (LLMs), natural language processing (NLP), automatic speech recognition (ASR), and text-to-speech (TTS), all leveraging NVIDIA software and pipelines. With natural interaction capabilities and a 92 percent accuracy rate for language understanding, FPT AI Engage can successfully solve simple customer queries and answer frequently asked questions, as well as complete more complex tasks such as telesales and payment reminders. This reduces repetitive workloads for customer service agents and improves overall efficiency.
By taking advantage of NVIDIA software and frameworks, FPT Smart Cloud has been able to expand the applications of virtual agents on an international scale, with the initial deployment in the Indonesian market. Currently, the AI vendor has implemented a total of 5,120 virtual assistants, processing more than 200 million interactions per month and serving 16 million end users. The adoption of NVIDIA cloud-native technologies, such as NVIDIA GPU Operator and Multi-Instance GPU (MIG) support in Kubernetes, allows optimal deployment at such a large scale. Trusted by more than 100 enterprises in 15 countries, FPT AI Engage has enabled, on average, a 50 percent increase in productivity and a 67 percent reduction in operating expenses.
FPT AI Engage enhances customer experiences using AI for virtual assistants and synthetic speech generation for model training.
Transforming Intelligent Call Center Operations in Consumer Finance
The capabilities of AI have attracted a plethora of banking and financial institutions in Vietnam, particularly Home Credit Vietnam. As one of Vietnam’s leading digital finance companies, Home Credit has always prioritized customer experience. Recognizing the importance of process automation and operational efficiency, Home Credit partnered with FPT Smart Cloud to deploy the FPT AI Engage solution in 2019, when AI was still a relatively new concept in Vietnam, demonstrating the company’s strategic vision.
After the first year of operation in 2020, FPT.AI Virtual Agent for Call Center supported Home Credit Vietnam to make more than 5,000,000 calls per month. This was scaled up, powered by NVIDIA, to 12,000,000 calls during peak hours, saving 50 percent of operating costs and achieving a 98 percent call success rate. In addition, the “virtual agent” of Home Credit Vietnam had an average customer satisfaction rating of 4.5/5.
The deployment is optimized using NVIDIA® TensorRT™ and served in NVIDIA Triton™ Inference Server with dynamic batching, saving up to 20 percent of high-performance computing resources for the same quality of model output.
So far, Home Credit Vietnam has put more than 100 use cases into applications, including information inquiries, self-service to lock or activate cards, automated customer surveys, and debt collection. As a result, these mundane tasks are offloaded to AI, and human agents have more time to handle critical customer issues.
One novel application of virtual assistants that can help convert a “cost center,” such as a customer service department, into a “profit center,” is a process called service to sales. FPT AI Engage increased Home Credit Vietnam’s service-to-sales volume by 2X.
These innovations have revealed the need for more human-like customer engagement in virtual assistants to boost the digital experience. This has turned the focus to the quality of AI-generated voices.
Synthetic voices are developed in tandem with virtual assistants, with a variety of tones, accents, and sentiments available. The type of voice deployed depends on the application. For instance, virtual assistants for general inquiries use more friendly and informative tones. AI voices for telesales involve more flexibility and emotions to better persuade prospects.
Advancing Speech Synthesis Models to Upgrade Conversational Quality
Recognizing the growing demand for more human-sounding, sentimental virtual assistants, FPT Smart Cloud aims to develop speech synthesis models that can produce new voices based on a few minutes of audio samples. The generated voices need to be high-quality, indistinguishable from the human voice, and communicate in multiple languages, even if the training inputs are in Vietnamese.
Given the nuances and emotional undertones of human language, the speech synthesis model often requires a vast amount of training data and long processing time to optimize accuracy and expressiveness.
Model training was previously executed on the NVIDIA A100 Tensor Core GPU. Typical training requires three servers with a processing capacity of 100 hours of voice data per day for a duration of 20 days. An upgrade to the NVIDIA H100 Tensor Core GPU is expected to handle more complicated model requirements and reduce processing time by at least 3X, or 7 days, with 2,000 hours of audio data.
With H100, the total process only requires one dedicated server. The speech synthesis model is ready in five days, a 4X efficiency improvement compared to the NVIDIA A100, as it processed about 400 hours of voice samples each day. The output model increased 100X in dimension and aptitude, generating a human-sounding voice that can seamlessly transition into 18 languages, including Vietnamese, English, and Indonesian.
Compared with the A100, the H100 is a step up in AI voice modeling, transcending language barriers and facilitating global communication on an unprecedented scale. Taking the leap forward in AI empowerment, FPT Smart Cloud is also utilizing the NVIDIA H100 for large language models to generate more accurate and flexible responses for virtual assistants.
“The protocols we have jointly established with FPT.AI have allowed us to create a fantastic, AI-based call center assistant. In doing so, Home Credit Vietnam has not only saved costs, but also been able to connect and help far more customers than we would with a traditional call center.”
Khang Pham Ngoc, General Director, Home Credit Vietnam
Harnessing the Power of NVIDIA GPUs to Foster Intelligent Workforce Development
AI has revolutionized various aspects of the business realm, particularly the structure of work. AI automation is taking over mundane, repetitive tasks with 24/7 operations and much higher efficiency. The workforce is transitioning to complex tasks that hinge on a broader range of skills, urging organizations to upskill and reskill employees to reach new working equilibriums as AI capabilities evolve and advance.
Powered by NVIDIA A100 and NVIDIA H100 GPUs, FPT AI Mentor is an AI solution that automates and personalizes employee training, based on the proven adaptive microlearning approach. It’s also a powerful virtual assistant that can support employees and resolve work-related inquiries, improving productivity and operating efficiency.
Augmenting the Workforce With AI
Long Chau is a leading pharmaceutical chain in Vietnam, owning more than 1,700 stores and employing 14,000 pharmacists, and has been growing rapidly over the past three years. With expansive operations across the country and strict industry standards, Long Chau has deployed FPT AI Mentor to ensure efficient, high-quality learning experiences for their employees and build a strong, competent workforce.
FPT AI Mentor uses LLMs built on NVIDIA DGX™ H100 and supported by NVIDIA NGC™ and PyTorch to generate and customize a variety of learning content from the business knowledge base. One hundred percent of Long Chau’s pharmacists receive daily five-minute quizzes that are personalized for the strong and weak points in each employee’s domain knowledge. Knowledge training in subsequent days focus on weak parts, combined with the spaced repetition method to maintain solid knowledge. This advanced approach has improved knowledge quality for the workforce by 55 percent while reducing required training resources by 80 percent.
Based on the assessment results, FPT AI Mentor generates individual knowledge graphs to indicate strong and weak areas of knowledge, ranking, and training progress, thereby creating an optimal development path. Managers can easily oversee workforce training quality with customizable, detailed reports on each department and each campaign.
“The rapid expansion of FPT Long Chau necessitates a highly skilled and dedicated team of pharmacists. Beyond traditional training centers, FPT Long Chau is leveraging FPT’s technological prowess by integrating AI into training programs to revolutionize pharmaceutical education. With FPT AI Mentor, our pharmacists have the flexibility to learn through personalized tests and quizzes directly from smartphones, ensuring that they are well equipped to deliver the highest quality of care to give better medical advice on a daily basis.”
Nguyen Bach Diep, Chairwoman, FPT Retail
Unlocking the Potential of LLMs in AI-Empowered Learning
Foreseeing the latest breakout of generative AI, FPT Smart Cloud has placed the focal point on generative AI development, starting with an upgrade of large language models. This initiative aims at diversifying learning formats to enable more efficient acquisition of business know-how for employees.
Training parameters grew by 10X in combination with more sophisticated learning objectives for the model and training data. NVIDIA H100 emerged as a cornerstone for AI and accelerated computing workloads. The H100 can speed up LLM training by 30X over the previous generation. The next-generation LLMs were built with NVIDIA NGC™, PyTorch, and two DGX H100 nodes. Application of NVIDIA-powered retrieval-augmented generation (RAG) workflows reduces how often pretrained models need to be updated and, at the same time, enables a wide and flexible range of tasks that can be done with the same pretrained foundation model.
LLMs enable FPT AI Mentor to generate situational questions and answers from the domain knowledge base. These simulations resemble real-life situations that employees might encounter on the job, allowing them to learn from practice and develop quick, effective reactions under realistic circumstances. Employees, however, can’t see the generated answers, and they’re exclusively used as a reference to evaluate employees’ responses. Generative AI and the advanced features its bringing to hands-on training is strengthening the competitive edge for FPT AI Mentor. Enterprise knowledge is embedded quickly into FPT AI Mentor using NVIDIA-powered RAG workflows, enabling the generation of relevant training content for specific enterprises.
Riding the exponential growth of AI, FPT Smart Cloud is utilizing accelerated computing from NVIDIA to transform traditional learning methods and elevate human capital development with tailored learning experiences, real-time feedback, and instant access to relevant information. These capabilities aren’t only improving knowledge retention and skill acquisition but also cultivating an agile, future-ready workforce.
