[The content of this article has been produced by our advertising partner.]
More than four million users now access OpenAI’s coding assistant Codex weekly, a recent uptake of one million users in two weeks proving just how quickly the competitive gap in developer tools is widening. Recognising this urgency, Amazon Web Services (AWS) has brought Codex and the latest artificial intelligence (AI) models to Amazon Bedrock. For CIOs and CTOs, this means capturing the fruits of agentic AI within existing security and governance frameworks.
At the What’s Next with AWS event on 28 April 2026, AWS unveiled its deepening partnership with OpenAI, explored the future of agentic AI, and outlined what businesses need to move from experimentation to real impact. The event also marked the rollout of a completely reimagined Amazon Connect and Amazon Quick, workforce solutions that demonstrate how agents are reshaping the work of businesses and builders.
As soon as Matt Garman, CEO at AWS, took the stage, he warned that many businesses are repeating a familiar mistake: simply adding agents to existing workflows is not enough to deliver transformational change. Instead, Garman urged organisations to rethink workflows entirely. "The real opportunity is when teams completely rethink what's possible by addressing questions such as how do the workflows change, how does the application change, how does the data change, and how does the UI change?"
For Garman, this shift does not spell the decline of employment but rather an elevation of human potential. “Across almost every single industry and every customer I talk to, agents are absolutely exploding and changing the way people do business. The nature of every job will change, but jobs aren't going away. The high-value work — understanding how to architect applications, solve customer problems, think about how technology fits together — is more valuable than ever.”
From Pilot to Production: Codex and GPT-5.5 Land on Bedrock
To facilitate enterprise transformation, AWS is innovating at every layer: from data centres and chips to the services that run on them. The most notable announcement, however, is that the latest OpenAI models, including the GPT-5.5 and GPT-5.4, will be available natively on Amazon Bedrock. For enterprise customers who have been hesitant to move beyond pilots due to security concerns, this development provides the confidence and green light to move forward in their agentic AI journey.
Two other limited preview offerings include Codex on Amazon Bedrock and Amazon Bedrock Managed Agents. By placing Codex on Bedrock, engineering teams can access the OpenAI coding agent within the AWS environments where they already operate at scale. Users can authenticate with their existing AWS credentials, process inference through Amazon Bedrock infrastructure, and apply Codex usage toward their AWS cloud commitments.
Meanwhile, Amazon Bedrock Managed Agents allows customers to deploy OpenAI’s "agent harness" - the runtime, environment, and inference API - directly inside AWS infrastructure. These agents possess persistent memory, specific skills, and a secure identity, delivering faster execution, sharper reasoning, and reliable steering of long-running tasks. They live natively within AWS, enabling access to EC2 instances or S3 buckets without exposing sensitive data to the public internet.
This integration of OpenAI's frontier intelligence with AWS's enterprise infrastructure is precisely what makes the partnership so compelling. Denise Dresser, Chief Revenue Officer at OpenAI, remarked, "When technology is moving this fast, making it easier to innovate in a trusted environment is incredibly important. Enterprise customers are no longer in the mindset of experimentation and pilots – they want to go full enterprise wide. They need powerful models in a trusted environment, so they can focus on building agents."
From Bot to Teammate: Amazon Quick Unifies Workspace
To ensure the power of agentic AI extends beyond software development and into the hands of every single employee, AWS also launched the reimagined Amazon Quick and Amazon Connect suite. These solutions are designed with different business functions in mind, so the people who keep businesses running day to day also benefit from agentic AI.
The new Amazon Quick desktop app is an AI work assistant that connects across local files, calendars, email, and business applications, learning from every interaction to understand what users do and how they work. By bridging silos across dozens of tools, including Slack, Microsoft Teams, Outlook into a unified workspace, Amazon Quick is no mere chat bot. Instead, it handles scheduling, builds deliverables, creates dashboards, and acts on your behalf.
Jigar Thakkar, Vice President of Agentic AI for Business at AWS, illustrated the vision behind Amazon Quick: “Quick connects to the systems you already use and builds a living map of your work, connecting the dots between people, projects, decisions, and actions. Quick knows you wherever your work happens. It learns what matters to you and your team, so one person's insight becomes the whole team's advantage.”
Setting up a meeting with a colleague would normally involve switching between Slack and calendar apps. Now with a single prompt, Quick checks the project background, scans schedule availability, sends meeting invites, and notifies colleagues - all in one go. By removing the tedious back and forth, early adopters such as BMW, 3M, and Mondelez have seen an 80% cut in production time for key processes and a 50% decrease in customer issue resolution time.
From Standalone to Full Suite: Amazon Connect for Every Function
On the other hand, Amazon Connect has undergone a huge transformation, with the name expanding to cover a full suite of products: Amazon Connect Decisions, Amazon Connect Talent, Amazon Connect Customer and Amazon Connect Health. The original product known as Amazon Connect will become Amazon Connect Customer, while Amazon Connect now stands for Agentic Business Solutions.
Colleen Aubrey, Senior Vice President of Applied AI Solutions at AWS explained this redesign : “The key insight is that if you constrain AI in a box, you won’t realise its potential. The real power comes when you carry context across the entire horizontal customer journey through your organisation.” The objective of the rebrand is to weave intelligence into enterprise operations as collaborative teammates and not isolated tools.
From Hesitation to Acceleration: Early Adopters Win the Race
The era of experimentation is over. With OpenAI models native on Bedrock, Codex accelerating software development, plus Amazon Quick and Connect transforming daily work, AWS has handed enterprises the blueprint for agentic AI adoption. The competitive gap continues to widen, but the infrastructure is ready and the security is proven. The discussion between AWS and OpenAI during the What’s Next with AWS event proves that it is time to integrate agentic AI into all facets of the organisation and turn frontier intelligence into competitive advantage.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。