8x8 adds live 13-language translation and multi-LLM choice to AI Studio
8x8 has updated its AI Studio platform with real-time voice translation across 13 languages and the ability for users to select from multiple large language models (Claude, Gemini, Grok, ChatGPT). These enhancements are designed to improve multilingual support and flexibility for contact center agents handling customer interactions. The updates also include one-click software connectors and tools to convert existing IVR setups to AI Studio call flows.
Key Takeaways
- Real-time voice translation detects language mismatches automatically, recording both original and translated audio for supervisor review.
- Users can now toggle between Claude, Gemini, Grok, and ChatGPT on a per-agent basis without changing the underlying platform.
- One-click connectors link AI agents to 15+ business applications, including HubSpot, Slack, Stripe, and Atlassian.
- A new conversion tool migrates existing legacy phone-tree IVRs into AI Studio call flows to modernize traditional customer service setups.
- Voice-led agent building now allows developers to use push-to-talk dictation with automatic filler word removal to generate instructions.
Why It Matters
This update shifts AI from a static assistant to a flexible, multilingual infrastructure layer by removing the 'vendor lock' associated with single-model platforms. By enabling real-time cross-language support natively, 8x8 directly challenges specialized interpretation services and reduces the latency typically found in third-party translation bridges. In the broader ecosystem, this move aligns with the industry transition toward 'Bring Your Own Model' (BYOM) architectures, where the value lies in orchestration and workflow integration rather than proprietary LLMs. Watch for whether 8x8's usage-based revenue—which grew over 70% in FY2026—continues to accelerate as these model-agnostic tools lower the barrier for enterprise-wide AI deployment.
Additional Context
The expansion follows 8x8’s pivot toward usage-based pricing models, an initiative that drove the company to its first full year of GAAP profitability in over a decade. Per No Jitter (May 2026), consumption-based revenue now accounts for 23% of 8x8’s total service income, up from 14% the previous year. CEO Samuel Wilson recently highlighted that generative and agentic AI are hollowing out traditional per-seat licensing, compelling vendors to charge per interaction or credit rather than per human worker. Technical performance in AI Studio was recently bolstered by the integration of OpenAI’s GPT-Realtime-2, which specifically improved the transcription and reasoning capabilities for accented and non-native speech. Per Business Wire (May 2026), this update provided the reliable transcription foundation necessary for the newly launched 13-language translation feature. This move is part of a broader trend where CCaaS providers like RingCentral and Genesys are increasingly competing on 'agentic' AI capabilities—autonomous systems that can execute multi-step workflows across diverse software stacks. Market data indicates that the demand for automated multilingual support is tied to rising operational expenses under traditional staffing models. According to CX Today reporting in June 2026, the average cost of a live agent interaction remains high at $17 or more, with agent attrition rates hovering near 31%. By automating linguistic bridges, 8x8 is positioning itself as a cost-offsetting tool for mid-market and enterprise centers that previously relied on high-cost bilingual staffing or external relay services.
Read full article at itbrief.com.au