Customer Service Automation · Sales Automation
FPT Smart Cloud needed more natural, multilingual virtual assistants and more scalable employee upskilling systems, but high speech-synthesis training costs, long processing times, and the need for personalized knowledge delivery limited how far those AI services could scale.
Reported impact
Across Engage and Mentor, FPT reports 3x lower data processing, 4x faster speech-synthesis processing, 100x richer speech-model output, 55 percent better agent knowledge quality, 80 percent lower training-resource requirements, 30x faster LLM training, and production-scale customer-service deployments that lifted productivity by 50 percent while cutting operating expense by 67 percent.
FPT Smart Cloud built two NVIDIA-accelerated application stacks: FPT AI Engage for customer-service virtual assistants and FPT AI Mentor for workforce development. FPT AI Engage combines LLMs, NLP, ASR, and TTS with NVIDIA software pipelines to automate inbound and outbound calling, IVR, telesales, reminders, and FAQ resolution. The deployment uses NVIDIA GPU Operator and MIG support in Kubernetes for multi-tenant scale, then optimizes inference with NVIDIA TensorRT and NVIDIA Triton Inference Server using dynamic batching. For model training, FPT moved speech-synthesis and LLM workloads from A100-based systems to NVIDIA H100 and DGX H100 nodes, alongside NVIDIA NGC, PyTorch, and NVIDIA-powered RAG workflows. That stack lets FPT build more human-like voices in 18 languages, personalize microlearning through AI Mentor, generate situational Q&A from enterprise knowledge bases, and deliver both call-center automation and staff training from a common accelerated AI foundation.
Customer calls and enterprise knowledge enter FPT AI Engage and Mentor, speech, language, and retrieval models run on Kubernetes-managed GPU infrastructure, TensorRT and Triton optimize the serving path, and the systems return virtual-agent responses, quizzes, summaries, or recommendations back into customer-service and workforce workflows.
FPT Smart Cloud, an NVIDIA Cloud Partner and leading AI provider in Vietnam, is using FPT AI Engage to accelerate their speech synthesis models by 4X.
The H100 can speed up LLM training by 30X over the previous generation.
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.
This advanced approach has improved knowledge quality for the workforce by 55 percent while reducing required training resources by 80 percent.
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.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Jun 26, 2026 MISA MISA needed to move from expensive, externally hosted LLM APIs to a self-hosted Vietnamese AI stack that could protect sensitive data, reduce latency and operating cost, and support many concurrent public-service and… $200,000 in monthly cost savings MISA AVA · NVIDIA A40 GPUs · NVIDIA DGX H200 
OpenAI Sample Dec 11, 2025 Podium Podium needed to help local businesses capture after-hours demand, follow brand-specific sales rules, and convert leads without forcing SMB owners to configure complex software or hire more staff. 30% increase in revenue on average ChatGPT Enterprise · GPT-5.1 · Jerry 
IBM Sample Feb 25, 2025 NDS Cognitive Labs NDS Cognitive Labs needed to help Clip support a fast-growing base of first-time digital-finance users at scale, answering questions about fees, setup, payments, and credit without relying on expensive manual support or… 28% annual increase in chatbot usage by Clip’s users IBM Cloud · IBM watsonx · NeuralSeek 