Amazon Connect’s AI Agent Upgrades & Deepgram’s Streaming Agents Integration At AWS re:Invent

At Amazon’s AWS re:Invent 2025, the company revealed that it upgraded its Amazon Connect offering, bringing a host of new features powered by agentic AI capabilities.

Amazon Connect, which serves as a complete AI-powered solution for delivering personalized customer experiences at scale, already helps businesses deliver automated voice experiences using neural text-to-speech in over 30 languages and automated speech recognition in over 25 languages.

Now, it’s bringing capabilities that can deliver natural voice interactions with advanced speech models,
and enable AI agents to:
 

• Understand, reason, and autonomously act on tasks across voice and messaging channels using deterministic and agentic experiences.

• Analyze conversation context and customer sentiment to complete tasks such as preparing documentation and handling routine processes.

• Provide AI-powered personalized product recommendations using customer history and real-time behavior to boost engagement.

• Be confidently deployed at scale with AI agent observability that brings transparency, visibility, and control over every customer interaction, allowing businesses to optimize performance and ensure compliance. 

“With advanced speech models from Nova Sonic, these agents deliver natural, human-like conversations, responding with the right pace, tone, and understanding across multiple languages and accents,” reads the release.

Customers using third-party automated speech recognition (ASR) and text-to-speech (TTS) solutions can now access Deepgram and ElevenLabs on Amazon Connect.

The idea of the upgrades is to enable businesses to leverage powerful AI capabilities that help them deliver seamless, personalized, and unified customer experiences.


Deepgram, which describes itself as the world’s most realistic and real-time Voice AI platform, announced native integration with Amazon SageMaker AI, bringing speech-to-text, text-to-speech, and voice agent capabilities powered by its streaming speech models.

It will also integrate its enterprise-grade speech technology with Amazon Connect and Amazon Lex, allowing customers to build and deploy voice-powered applications with sub-second latency within their AWS environments.

Meanwhile, Deepgram’s clients will be able to deploy real-time speech capabilities across AWS services.

“Integrating Deepgram's advanced speech technology with Amazon Connect enables organizations to build voice interactions that understand context and respond with appropriate pace and tone, transforming automated interactions into opportunities for deeper customer relationships,” said Pasquale DeMaio, VP of Amazon Connect at AWS.

Amazon’s AWS re:Invent 2025 will be held from December 1 to 5 across multiple venues on the Las Vegas Strip, spanning Mandalay Bay, MGM Grand, The Venetian, Caesars Forum, and Encore.

The event will see Amazon and AWS (Amazon Web Services), its cloud arm, unveil and launch a series of new products, services, and announcements regarding agentic AI, cloud infrastructure, and foundation models.

As part of the event, Amazon announced AWS Interconnect - multicloud, a new solution that enables simplified multicloud connectivity between AWS and other major cloud service providers. The first partner on this first-of-its-kind purpose-built network interoperability product initiative will be Google Cloud.

Currently in preview, the move is being described as “a step toward a more open cloud environment,” and Amazon plans to add more cloud providers soon. 

Do you think these moves will help Amazon revolutionize the AI agents and AI-voice interaction market?

Let us know in the comments below!