SoundHound AI targets 90% OASYS migration following LivePerson acquisition deal

The integration of LivePerson’s chat-based workflows into the OASYS platform signals a shift from specialized voice tools to comprehensive omnichannel AI agents. By combining proprietary voice-native models with third-party LLMs from Amazon Bedrock and Hugging Face, SoundHound is positioning itself as a production-grade alternative to horizontal orchestration platforms. This move directly addresses the growing demand for voice-enabled customer service in the automotive and retail sectors, where noise-resistant, low-latency interactions are critical. The strategy hinges on converting LivePerson’s legacy chat customers to a unified voice-and-text stack. Watch for updated 2026 revenue guidance once the transaction officially closes to gauge the immediate financial impact of the combined enterprise footprint.

SoundHound AI has been building its OASYS orchestration platform as a differentiator against horizontal conversational AI vendors. The company reported that OASYS integrates proprietary voice models with third-party LLMs from Amazon Bedrock and Hugging Face, creating a multi-model architecture designed for production deployments in automotive, restaurant, and retail verticals. This positions SoundHound against broader agentic AI platforms from Salesforce, Microsoft, and Amazon that lack specialized voice-native processing. The OASYS platform is central to the company's pitch that vertical AI agents outperform general-purpose chatbots in latency-sensitive environments like drive-through ordering and in-car assistants.

The LivePerson acquisition fits into a broader wave of consolidation in conversational AI. In early 2025, LivePerson reported that its enterprise customer base had been declining as brands shifted budgets toward generative AI platforms, with its stock falling sharply on weak forward guidance. SoundHound's $43 million all-stock bid represents a significant discount to LivePerson's historical valuation, which peaked above $8 billion in 2021. The deal structure reflects how legacy chatbot vendors have lost pricing power as large language models commoditized basic text-based customer service. SoundHound CEO Keyvan Mohajer has framed the acquisition as a way to add enterprise chat workflows to its voice-first platform without building from scratch, targeting LivePerson's remaining installed base of Fortune 500 accounts.

On the technical side, SoundHound's Polaris voice model has been benchmarked for sub-500-millisecond response latency in noisy environments, a key requirement for automotive and quick-service restaurant deployments. The company announced partnerships with major automotive OEMs including Stellantis and Hyundai for in-vehicle voice assistants running on its platform. Competitors in the voice AI space include Cerence, which serves automotive OEMs with its own proprietary models, and Baidu's DuerOS for the Chinese market. SoundHound's Q2 2025 revenue of $61.9 million, up 45% year over year, placed it ahead of analyst consensus estimates of approximately $57 million, suggesting that the OASYS migration strategy is gaining traction with existing customers even before the LivePerson integration begins.