At its annual re:Invent 2025 event in Las Vegas, AWS announced a major update to its cloud-based customer service solution, Amazon Connect, officially introducing "Agentic AI" capabilities and adopting advanced...Nova Sonic speech modelThis allows AI to not only engage in conversations, but also to proactively reason and execute tasks.
In addition, at the infrastructure level, AWS has also made a rare announcement of a collaboration with Google Cloud to launch a preview version of AWS Interconnect – multicloud, which will solve the long-standing problem of multi-cloud network connectivity complexity faced by enterprises.
Amazon Connect Evolution: From Automation to "Agent-Based" Collaboration
The core of this Amazon Connect update lies in the introduction of self-service capabilities from Agentic AI. By integrating Nova Sonic's advanced speech model, the new AI agent can understand and respond fluently to customer needs across languages with a more natural, human-like tone, rhythm, and accent.
Furthermore, AWS emphasizes that these AI agents possess both "reasoning" and "acting" capabilities. While conversing with customers, the AI can analyze semantics and emotions, and proactively assist human customer service representatives in completing background tasks such as document preparation and handling routine processes, thereby achieving true human-machine collaboration: AI handles complex backend processes, allowing human customer service representatives to focus on building customer relationships and handling complex situations.
In addition, for customers who are already using third-party voice technologies, Amazon Connect now also supports solutions from providers such as Deepgram and ElevenLabs.
To ensure compliance and compliance, AWS has also launched the AI Agent Observability feature. Enterprises can use a transparent interface to see what AI understands, what tools it uses, and how it makes decisions, thereby optimizing AI interaction performance and building user trust in AI.
Deepgram integrates SageMaker AI to achieve sub-second voice processing.
In addition, Deepgram, a voice AI company, also announced at the conference that it will bring its advanced voice capabilities to the AWS ecosystem. Deepgram's streaming speech-to-text (STT), text-to-speech (TTS), and voice proxy capabilities will be directly integrated into Amazon SageMaker AI, Amazon Connect, and Amazon Lex.
Through this integration, businesses can build real-time voice applications with sub-second latency within the secure environment of AWS, improving the responsiveness and fluency of automated interactions.
In an unprecedented move, AWS has partnered with Google Cloud to launch AWS Interconnect – multicloud.
In terms of cloud infrastructure, AWS announced a partnership with Google Cloud to launch a preview of AWS Interconnect – multicloud and jointly developed new open specifications for network interoperability.
This solution combines AWS Interconnect and Google Cloud Cross-Cloud Interconnect technologies, enabling enterprises to establish dedicated, high-bandwidth private connections between the two cloud providers. Previously, when connecting different cloud workloads, enterprises often faced the dilemma of insufficient bandwidth guarantees when using public networks, or the overly complex nature of establishing private connections.
This fully managed service allows businesses to quickly configure cloud-to-cloud connectivity via the AWS management console or API, significantly reducing the burden of managing physical devices or virtual routers. AWS has also released the relevant open API suite through GitHub to encourage more service providers to join this interconnected ecosystem.
Cross-cloud services have already been operating before, and the next step is to integrate more cross-cloud applications for development.
In fact, this is not the first time Google Cloud has integrated cross-cloud services with AWS; the idea was first proposed back in 2020.BigQuery Omni cross-cloud search functionalityThis allows users to search for files across Google Cloud and AWS cloud service platforms, and will later be integrated with Microsoft Azure for the same service.
Amid the current trend of AI applications, Google Cloud also announced the launch of [a new technology/service] this year.A2A (Agent 2 Agent) ProtocolThis enables its AI agent services to interact across clouds, and Microsoft subsequently joined in the integration. AWS also recently announced its support for this protocol, allowing more AI services to operate across clouds.
AWS's announcement of further cross-cloud integration with Google Cloud suggests that the two companies will integrate more cross-cloud applications in response to AI development trends.
