Slowly but surely, Amazon’s AWS cloud computing unit has become a major player in the call/contact center space with its Amazon Connect cloud-based (and AI-centric) contact center service, which launched back in 2017. Today, companies like Air Canada, Dish Network, and U.S. Bank use the platform for their customer service needs. At its annual re:Invent conference in Las Vegas, the company has now announced a number of updates to Connect which, unsurprisingly, focus on AI, powered by the Amazon Q platform.

“When we first came out, we were really a voice-only solution that focused heavily on bringing AI to the contact center [with] scalability, security — the things that are our calling cards for AWS. And pretty quickly, we were able to add more features and get to a bigger feature completeness,” Pasquale DeMaio, vice president and general manager of Amazon Connect at AWS, told me. “Now, we offer channels across everything from chat, email — coming out as we speak — and also SMS, WhatsApp, Apple Messaging for Business.”

DeMaio stressed that AWS built Connect as an end-to-end solution that is now in use by over 14,000 external customers, as well as Amazon.com itself.

Given the contact center context, most of the new features focus on how Connect customers can more easily build AI-powered self-service workflows that can handle many of the more routine customer service tasks. Originally, AWS used Q in Connect mostly to help guide agents through their customer interactions. Now, businesses can use the service to build customer-facing self-service experiences as well.

To ensure these external-facing conversations don’t go off track, AWS allows businesses to set customized guardrails to keep the conversations on track, reduce hallucinations, and help the bots adhere to a company’s preset policies.

Ideally, all of this frees the human agents to focus on higher-value and more complex interactions, DeMaio noted. And talking about those human agents, Connect is also launching new AI-powered agent evaluation tools that the company says will “enable customer service managers to easily spot performance trends, enhance training, and help improve overall service quality.”

What’s maybe even more interesting here, though — and something you may see pop up as a customer calling into a call center soon — is that AWS is trying to use all of this data and generative AI to help businesses be more proactive in their customer interactions.

“I think the best customer service is often proactive, not always, but often proactive,” DeMaio said. “And it’s been sorely lacking over time, because it has been hard […] but if it’s gotten right, it really can be terrific.”

With this release of Connect, the team built tools to help businesses track what’s happening with customers in real time (maybe a flight is delayed, a package is stuck in transit, or a subscription is about to renew), segment them into different groups, and then reach out proactively on the most appropriate channel. Ideally that’s a better customer experience, but it also reduces the number of times customers have to contact the company, which will likely save the business money in the long run.

All of this is typically powered by integrating a number of disparate systems with Amazon Q Business. Sometimes that also goes the other way around, with third-party customers building AWS Connect into their contact center solutions. Salesforce, for example, is launching the “Salesforce Contact Center with Amazon Connect” today, which integrated Amazon Connect’s core capabilities with unified routing into Salesforce’s CRM solution.

“Companies can now use a single routing and workflow solution for their Amazon Connect and Salesforce channels to intelligently deliver calls, chats, email, and cases to the right self-service or agent interaction,” AWS explains.

It’s worth noting that AWS is aware that not every Connect customer is ready to use generative AI just yet. “When I talk to customers in the real world who are trying to do this, their big thing is: please stop trying to ram [generative] AI down my throat for every solution,” DeMaio said. “We want to help you go at your own pace and do it the right way for your business, and be able to use it for the things where it’s useful but rely on other technologies that already work great. And I will say, there’s even situations where a touch tone is still as good or better than voice, like if you ask me to enter my credit card number.”