Dec 04 2025
Artificial Intelligence

AWS re:Invent 2025: AI Agents Want to Be Your Teammate

But first, you’ll have to stop treating it like just another tool, according to speakers at the annual tech event.

At AWS re:Invent in Las Vegas, speakers argued that the next big productivity leap will come when companies stop making workers hop between systems and instead hand them an “intelligent teammate” that understands context, takes action, and stitches workflows together.

“It’s not about artificial intelligence as a tool,” Pasquale DeMaio, vice president and general manager of Amazon Connect, told a packed room. “We need to make AI human-centered — bringing AI to meet you where you’re at and meet your workforce where they are.”

DeMaio didn’t sugarcoat the state of play when it comes to AI projects. “We’re hearing that 42% of these projects are failing,” he said, while acknowledging that other estimates run even higher. But where projects succeed, the benefits can be dramatic: “When these human-plus-AI integrations work, they’re seeing a 70% boost in work completion.”

Click the banner below for deeper insight into IT modernization.

 

How Amazon Quick Suite Automates Work Flows

Speakers drew contrasts between the current, fragmented corporate day, with workers wasting time hunting through inboxes, dashboards, and multiple apps for a single answer, and a future where AI agents orchestrate those sources. Jose Kunnackal, director of Amazon Quick Suite for AWS, said individual AI solutions have proliferated, saving pockets of time but creating “AI sprawl.” Teams end up managing dozens of disconnected assistants, each with its own login, governance, and data silo.

Quick Suite, Kunnackal said, is AWS’s answer to that sprawl, a “unified workspace that brings together context across all of these information sources” with agent roles for insights, research and automation. Instead of toggling between Microsoft SharePoint, dashboards, ticketing systems and email, users can ask a single agent to pull and reason across that context, and then take action — from drafting an email to scheduling a meeting.

For example, he said, imagine a manager asking for “a Q3 performance update.” The agent can gather everything required for that task: pull financial or operational documents from SharePoint, grab ticket-volume metrics from systems such as ServiceNow, reference relevant meetings or milestones from the Outlook calendar, and then summarize all of that into a slide deck. It can even package the deck into an email and send it to a specific colleague.

Meanwhile, that same natural-language request can be turned into a reusable workflow. Instead of the manager having to repeat the instructions every quarter, the agent can run the entire multistep process automatically — on a schedule, on a trigger, or on demand. In this way, Kunnackal, said, Quick Suite can turn an ad-hoc conversational command into a fully automated business process, letting knowledge workers hand off repetitive reporting cycles to an AI agent.

Click the banner below to learn how enterprises are unlocking artificial intelligence’s potential.

 

How AstraZeneca Is Using Quick Suite

Vaishali Goyal, senior director of R&D IT at AstraZeneca, said clinical researchers had been spending lots of time chasing research and market signals. Quick Suite let the company automate collection and synthesis so leaders and practitioners receive “concise and consistent” insights. AstraZeneca has piloted the tech in hematology and plans a broader roll-out across therapeutic areas in 2026.

Speakers returned repeatedly to guardrails. Success isn’t having AI “take over and write the work slop you’re hearing so much about,” DeMaio warned. Instead, properly integrated AI should improve both efficiency and quality — he cited a 25% improvement in task completion and “40% higher quality of work outcome” when human and AI operate cohesively, according to AWS’ own research.

Soeakers underscored two contradictory truths companies face. On one hand, the productivity gains to be won with generative AI are significant: multipliers of output, faster decision cycles and fewer duplicated efforts. On the other, execution is fiendishly difficult, with projects are failing at alarming rates.

“If you build AI as something that’s amplifying human capabilities … you’re taking away the things that are mundane, boring, and don’t bring real value,” DeMaio said.

Close

New Workspace Modernization Research from CDW

See how IT leaders are tackling workspace modernization opportunities and challenges.