AWS’s commitment to AI agents in customer service: Some contact centers attempt full automation. Some people use AI agents just to assist humans. Most of the time, the two come together somewhere in between.
This is the concept behind the bundle of Amazon Connect features we released today, including contact center agents, agent observability, and orchestration tools.
Among the 29 customer service AI agents and related tools on display by AWS (most live, some in preview) were agent assistance tools that include real-time conversation analytics that can suggest answers, create complete documentation, and set up digital workspaces for each persona, such as an agent or manager. Observability features include evaluation and monitoring tools for AI agents and test environments.
The latest AWS releases also include tools for AI agent orchestration. This feature set includes connections to third-party speech-to-text agents from Celebrities and Deepgram, which can be orchestrated within Amazon Connect.
Currently, customer service vendors like Salesforce and ServiceNow (including Qualtrics) are competing for control of orchestrating their own AI agents and those of other companies. The situation is fluid, and AWS doesn’t claim to be “rigorous” about which AI or orchestration tools customers should use, said Pasquale DeMaio, vice president of Amazon Connect at AWS.
“My honest opinion on this is there’s going to be a lot of winners here,” DeMaio said. “It’s a big place. Some people are going to make great bets that work out, and some people are going to make bad bets that don’t work out. Either way, there’s going to be a lot of bets. … I don’t think it’s going to be one winner-take-all, but I think we’re going to be leading the way.”
AWS has also integrated many of its contact center AI capabilities within its Salesforce Service Cloud environment. Joint customers can enable human agents to access AWS’s generative AI agent assistance, predictive recommendations, and automated workflows from within Salesforce. Amazon Connect now supports WhatsApp, a popular channel internationally, especially in regions such as India, Latin America, Indonesia, and the Philippines.
AWS integrated environment
Opus Research analyst Ian Jacobs said that while AWS isn’t a pioneer in the customer service AI agent space — several other companies have already deployed their own agents, including Salesforce, ServiceNow, Nice and Genesys — it doesn’t need to be. This technology is still in the early stages of implementation. Many contact centers are just getting started.
In fact, Jacobs noted that some contact centers are still preparing their environments for AI, combining hardware and cloud software to launch their first contact center as a service instance. AWS can provide contact centers with scalability, reliability, and a unified cloud environment for agent AI without having to set up numerous add-ons from other vendors for features such as agent observability and management, voice channel integration, and Amazon Bedrock machine learning support for your knowledge base.
Additionally, as Jacobs notes, AWS gives developer-centric shops the ability to “ship, test, and debug AI agents without inflating their AI bills.”
“They have no first-mover advantage… [but] Boring reliability is a huge advantage,” says Jacobs. “Now everyone has an AI agent. The ways AWS can differentiate are in routing, flexibility, transparency, and configurability.”
AWS’s Amazon Connect agent was released in conjunction with the AWS re:Invent conference held in Las Vegas from December 1st to 5th.
Don Fluckinger is a senior news writer at Informa TechTarget. He is responsible for customer experience, digital experience management, and end-user computing. Any tips? send him an email.
