Built from the inside out: How AWS Professional Services became the first frontier team
AWS Professional Services (AWS ProServe) has compressed engagement times from months to days, not by adding artificial intelligence (AI) tools to the existing process, but by reinventing the way we deliver internally and externally. The shift reflects what my colleague Swami Sivasubramanian pointed out in How Frontier Teams Are Reinventing AI-Native Development: the real productivity gains come from rethinking how software is built, not from putting AI into existing workflows.
In this post, I'll share how AWS ProServe became an edge team, the processes that enabled it, and what your engineering organization can take away from our experience.
My partner has already done it
Building a boundary team is something every organization can do. For customers looking for help getting up to speed, AWS ProServe is a partner whose consultants are already involved in the advancement of AI in the way they work every day.
AI-native development is moving at a pace traditional consulting methods were never designed for. A task that used to take months is compressed into days, and the rhythm changes accordingly: tighter loops, faster feedback, more decisions made in the course of the plot. Helping the customer to work on time requires consultants who know which decisions can move quickly, which require careful human judgment, and how to keep quality high when execution is fast. That sense comes from doing the work.
Our pilot mimicked what Swami described: free advisors from outside of coding (scripts, linking, status reporting, iterative scaffolding) that used a large part of all interactions. This allows human judgment to be focused on where it actually drives results. So we did what border teams do. We invested in the agent context, reorganized work around what agents do well, and stopped treating AI as an assistant. We started treating it as a foundation.
Our team of pathfinders: APEX
Swami's blog describes the three approaches Amazon's teams take to native AI development: the pathfinder campaign, the structured sprint, and in-situ testing. AWS ProServe started as a pathfinder.
Our Agentic AI ProServe Experiences (APEX) team had one mission: to redesign how ProServe delivers. APEX built the ProServe Delivery Agent, multi-agent system requirements, architecture verification, implementation, security updates, testing, and deployment. A managing agent organizes specialized sub-agents at each stage of the life cycle.
Delivery Agent is how ProServe works AI-DLCi AI-Driven Development Lifecycle. AI-DLC was developed by AWS platform teams, developed and refined through hundreds of customer workshops. AI development is fundamental. AI-DLC is a process built for AWS to drive it through the full delivery lifecycle, for us and our customers.
APEX has proven the model in its production load. The Delivery Agent now works alongside human advisors in global negotiations, and proven APEX patterns become the default delivery move across ProServe. This is not a pilot. It's how we deliver on scale.
How we reorganized the delivery movement
ProServe's standard collaboration is used to follow a regular consultation rhythm: access to long documents, architectural decisions discussed in conferences, implementation of sprint cadence, testing and safety at section boundaries. Each handoff introduced lag, and each artifact was written for human use only.
The redesign changed every step. Requirements have been moved from prose to structured documents that both people and agents can read, becoming a source of truth rather than a product. Building standards and lessons learned from past interactions were compiled into guidance files that agents continually draw upon. The implementation has changed from contributors working sequentially on tickets to coordinators feeding well-planned tasks to multiple agents in parallel. Security testing and review has moved into the build loop, with agents verifying output on site and remediation before any human review begins. Status reporting and coordination largely disappeared.
The net result: continuous flow, with human judgment focused on priorities, validation, and higher decisions.
Creating a delivery agent using Delivery Agent
APEX builds a Delivery Agent using native AI processes that it offers to customers. A feature request goes into the system. Agents generate structured tickets, generate code, and perform automated testing through our GitLab-integrated DevOps pipeline. In the review and approval of the person, the change is implemented.
People manage judgment: setting priorities, strengthening quality, authorizing high-level decisions. Agents manage the scaffolding. Lower resolutions operate independently. People's gates focus where judgment is important.
As delivery teams across ProServe embrace Delivery Agent in collaboration, they provide learnings and, sharpening every project. That's how Amazon built it. We use our own products, see what breaks, and fix it.
Five ways to make this work
Five practices from Swami's blog now describe how we use AI-DLC within ProServe:
Slow down to speed up. Frontier teams invest before they accelerate, build an agent core and set up a practice before the integration of speed. APEX made that investment once, so we transfer muscle memory directly rather than asking each customer to start from scratch.
Invest heavily in agent context. Guide files and architecture standards are first-order artifacts in every interaction. The more context there is, the safer the agent can use them.
Feed agents instead of guarding them. Builders maintain a strong backlog of well-planned tasks and run multiple agents in parallel, updating the output in parallel.
Use the specs as the source of truth. Certain driven improvements are automated workflows. Details are not documents. They are contract agents who build against them.
Shift to check left. Agents validate locally and self-correct before output reaches human review.
Delivery of AI to customer environments
In customer interaction, the Delivery Agent works together with the human coordinators. Together they work across the entire lifecycle, from planning to implementation, against the business outcomes the customer has chosen. The governing principle: humans provide purpose, AI creates, humans verify.
Customers retain the choice of basic models and can extend the system with their own data and tools.
What we learned by going first
Balancing is not an option, but you don't have to start from zero. Teams need time to build trust in what agents are good at, break down complex work into verifiable tasks, and redesign artifacts for AI use. We transfer that muscle memory directly during the joint, shortening the curve.
The workflow is constant. Empowering tools. We use Kiro, Amazon Bedrock AgentCore, and Strands, but the stack is not what creates the productivity advantage. Tools only integrate when workflows are reorganized around them.
Match the results. Traditional consulting costs time and materials, promoting duration over impact. We have moved to a fixed price partnership that is consistent with the business results used in production. If the business model matches the needs of the customers, everything else follows.
Real results
“We adopted Amazon Application Recovery Controller's (ARC) new functionality to streamline our multi-region resiliency approach. Regional Switch replaced custom failover orchestration with declarative systems that include scaling, database replacement, and DNS routing across our services in parallel. Kiro with AWS Professional Services compressed 6-week code delivery, accelerated AC delivery and quality enforcement constant in all deliverables of our region performed on time and we were able to work in our second location.Matt McKeever, CTO Infrastructure and Operations, LexisNexis Legal & Professional.
Getting started
Workshops: AWS Solutions Architects conducts AI-DLC workshops, two- to five-day engagements that demonstrate native AI development against your stack. Hundreds of customers have already participated.
Production agreements: When you're ready to take business use cases to production, AWS ProServe steps in. Our Consultants and Delivery Agents are embedded with your team to deliver productive results while building organizational capacity to support and grow the practice. Finally, you have applications in production and trained internal champions ready to move forward.
“Our Solutions Architects have been at the forefront of this change, collaborating with customers at AI-DLC conferences to rethink how they build software. When teams experience AI advancements firsthand, they don't want to go back. AWS Professional Services takes that momentum and leverages it, on a big scale.” Shaown Nandi, Vice President, Technology, AWS.
Many organizations have results waiting to happen and engineering teams ready to work differently. The method is not more testing. Committed to working with a team that has already proven the way to produce their work.
Contact your AWS account team or visit the AWS Professional Services webpage to start delivering productivity results quickly.
About the author
