Breaking News

Mihup and Qualcomm collaborate to advance secure on-device voice AI for BFSI

Mihup has announced a collaboration with Qualcomm Technologies to develop and commercialize multilingual, enterprise-grade Voice AI solutions tailored for the Banking, Financial Services, and Insurance (BFSI) sector. The joint effort focuses on optimizing voice intelligence to run directly on on-device AI platforms, reducing reliance on cloud infrastructure while enhancing privacy, performance, and scalability.

The partnership aims to address key challenges faced by BFSI organizations as they accelerate digital transformation. Traditional cloud-based Voice AI deployments often encounter issues related to latency, bandwidth limitations, infrastructure costs, and data security. By enabling AI workloads to operate directly on devices powered by Qualcomm’s AI processing capabilities, the collaboration seeks to deliver lower latency, improved data governance, and stronger privacy protections for regulated industries.

Hybrid Architecture for Cost Efficiency and Performance

Mihup’s Voice AI contact center solution will adopt a hybrid architecture. High-volume tasks such as speech-to-text processing and real-time agent assistance will run locally on Qualcomm’s Neural Processing Units (NPUs), while more advanced functions can scale via the cloud when necessary. This approach is designed to significantly lower operational costs and reduce dependency on constant cloud connectivity.

According to Mihup’s internal analysis, shifting intensive speech workloads to on-device processing can reduce total cost of ownership by up to 78 percent, depending on scale and usage. The companies state that this model enables faster response times, real-time conversational insights, and enhanced customer engagement—particularly important in linguistically diverse markets such as India.

Built for Regulated Industries and Global Expansion

The solution is engineered to support multilingual customer service, sales, and collections functions, reflecting natural conversational patterns across regional languages. By processing sensitive data directly on devices, the architecture aims to ensure greater compliance and data sovereignty for financial institutions.

With commercial deployments underway in India serving as proof of scale, the collaboration is expected to validate on-device AI as a viable model for enterprise Voice AI adoption. While initially designed for India’s BFSI ecosystem, the companies plan to position the solution for expansion into other developing markets and regulated industries worldwide.