AI and brokers have been large focuses at AWS re:Invent 2025, with the cloud big unveiling a number of recent companies and platforms to supercharge capabilities and encourage wider utilization.
But with a lot hype and hypothesis across the know-how, how reasonable can these lofty objectives truly be? I spoke to among the main minds at Amazon and AWS to seek out out extra.
Empowering with AI
“We want to empower every application developer to be an agentic developer who builds agents that are reliable, trustworthy, secure and making it super easy to build,” Swami Sivasubramanian, Vice President, Agentic AI at AWS, starts off our briefing by boldly declaring.
“Developers are no longer constrained by programming languages and syntax, so the speed of development is going from years to months, and months to weeks.”
Sivasubramanian notes how, “every job discipline is fundamentally getting turbo-charged with AI agents” – and this includes the company’s new “frontier agents”, which look to handle monotonous or time-intensive tasks so a human employee can focus on most engaging work, such as writing code, and Kiro, an agentic IDE for vibe coding, which Sivasubramanian says has had an “amazing reaction” from hundreds of thousands of users already signing up ahead of its now-open general availability.
“The great news is this is just the start,” he says, “in many ways, the world of software development is going to fundamentally change”, noting how the role of a full stack developer will move from being someone who simply wears many hats, to someone who can work with agentic teammates to create the future – what the company calls “renaissance developers”.
Looking internally is another key motivator and inspiration, he adds, with the scale of Amazon itself, covering everything from ecommerce to cloud computing to rockets, means its internal workflows can be a good initial use case.
“The needs are different, but the fundamental problems are the same,” he says, “I do think this is going to fundamentally change everything,” he concludes, “AgentCore is going to be the core building block for everyone – what AWS did for the IT industry, Agentcore will do for every agent out there.”
The success of Kiro, released in July 2025, is a nice shock for Sivasubramanian, as he notes how startups and huge firms alike are actually getting concerned in AI transformation, and it’s a superb signal for AWS’ future product growth too.
These new releases should not simply designed for in the present day’s employees, he notes, but additionally the subsequent era as properly, who will probably strategy work in an AI-native, AI-first approach.
“We envision a world in AWS the place we have now the inspiration for billions of brokers,” he says, “and for me, that alone is an understatement – we want every employee out there to have an agentic teammate that amplifies their capabilities and solves the undifferentiated, drudgery of certain tasks.”
AI with trust
The crucial question for many users then, is how do we ensure there is a human presence, and responsibility isn’t solely handed over to AI tools.
“Agents are not going to reach their fundamental promise unless they are trustworthy,” Sivasubramanian says, noting how the answer is making certain grounding in stable, verifiable sources, not simply endeavor automated choices with out correct consideration – together with AWS’ Agentcore interception instrument, which might alert a human consumer to examine and confirm if wanted.
But customization ought to nonetheless be an essential focus, he notes, as customers will all the time need various things, and be working at totally different ranges of experience. Technology like the brand new AWS Nova Forge, which lets organizations get pleasure from beforehand unheard-of ranges of customization, generally is a gamechanger without having to interrupt the financial institution.
Also at AWS re:Invent 2025, I spoke to Rohit Prasad, SVP and Head Scientist, Artificial General Intelligence at Amazon, who outlined the “persistent trade problem” of foundation models, and the “stark actuality” of at-scale functions making an attempt to run them.
“This is making everyone yearn for a model that’s an expert in their domain, their business, their knowledge sources and so on,” Prasad says. “Intelligence is not monolithic…the ultimate utility is in specialized domains – every organization is trying to differentiate their business…and this is where AWS has just been world-class in making technology very accessible.”
And Sivasubramanian agrees, noting how “this concept of open training being accessible – if you look back at the history of technology, all the essential technologies used to start with only experts using them, before becoming easier and easier, and expanding so everyone can use them.”
So wanting ahead, it’s clear that AWS sees brokers as the subsequent large step ahead within the working world – with CEO Matt Garman declaring the know-how will likely be “bigger than the internet or cloud”.
I ask Sivasubramanian what he thinks of this, and whether or not it’s a viable outlook.
“Agents are going to fundamentally transform how we live and work, of that I’m a true believer,” he laughs, “but what that means is that in order to achieve that future, you also need to build all the right foundational infrastructure and the frontier agents – and that’s exactly what we’re doing at AWS.”
“Many tasks are going to get automated, but the underlying jobs, and what we do, and what we’re trying to accomplish, are not going to go away,” he says, “we are going to be able to do them faster and better, and we will move on to solving other challenges.”
“The era of agents is going to take time,” Prasad says, “I do believe that we will live in an Internet of Agents, which will do tasks on your behalf, communicate on your behalf…but for agents to really work for you, you have to be able to trust them – and trust comes from reliability.”
“If you look back at the history of major changes like the Industrial Revolution, or the beginning of the Internet or Cloud…certain things did change – but what people had to do did not change,” Sivasubramanian concludes.
“You’re going to see unparalleled productivity in terms of what you do, and we will no longer be constrained by our ability to build complex things, we are only going to be constrained by our creativity – and that’s what makes this really exciting.”