AWS Forward Deployed Engineering
Turn AI into business impact, with the engineers who built AWS AI. AWS FDE brings the expertise to change the work and system to make it last.
What is AWS Forward Deployed Engineering?
Every CEO is being asked to show real value from AI, and faster. The tools are everywhere. The results aren’t. That’s because real work is harder than a pilot: the know-how lives in people’s heads, older systems weren’t built for agents, and every use case starts from zero. Value comes from redesigning the work end to end, on a foundation that grows with each use case, with a team that can keep extending it.
That’s what AWS Forward Deployed Engineering (FDE) is built for. Together, you take a priority business problem from idea to impact, on an agentic system you keep using when AWS FDE steps back.
Your business, our experts, and a system you keep using
Discovery agents map your systems and data, the real rules that live in people’s heads and legacy apps, and the language your teams use. The result is a living model of your business that shows where to start and gives every agent the same picture.
The engineers, scientists, and product managers who built frontier AWS AI services and run AI in production at AWS scale work side by side with your engineers. They redesign the work, lead the change, and prove the outcome, and your engineers learn by building with them.
The agentic system AWS FDE builds with runs in your own environment: discovery, a workbench where AWS FDEs, your engineers, and agent crews build from proven accelerators on the latest AWS AI services. When the FDE team steps back, you keep using the system and your team runs it.
Speed: results in weeks. AWS FDE never starts from a blank page. Discovery turns your goal into a clear outcome in days, agent crews build alongside your engineers from accelerators already proven in production, and the work runs on your real data and systems from day one.
Scale: each use case builds on the last. Anyone can move fast on one use case. With AWS FDE, the living context, controls, and code from the first use case carry into the next, so new agents use the existing foundation, and value compounds as you extend across your business.
Control: yours to run and extend. Everything runs in your environment, under your identity, network, and security controls, with guardrails, evals, audit trail, and cost controls built in from day one. It’s open to the models and tools you already use.
Customers who’ve built with AWS FDE
Marsh
Now spanning three lines of business, with dozens of workflows planned. Core workflows were rebuilt with agentic AI, inside Marsh's own environment and alongside its Oliver Wyman consulting team. Throughput rose 53%, and code reviews run 56% faster.
“We set out to build a firm that reinvents itself continuously. AWS forward deployed engineers embedded with our teams and rebuilt core workflows with agentic AI inside our own environment, working alongside our Oliver Wyman consulting team, turning that ambition into capabilities our people use every day. The engagement gave us both the speed and a repeatable way to keep going.”
- Paul Beswick, SVP & Chief Information and Operations Officer, Marsh
Samsung Life Insurance
Samsung Life Insurance had four affiliates, each building its own AI point solutions from scratch. In 45 days, the FDE team built an end-to-end governed AI-DLC harness and control plane. New agents now join by registration and policy assignment.
Rio Tinto
Rio Tinto Iron Ore runs one of the world's most complex mine-to-ship supply chains on a 30-year-old Oracle system. That system has more than 5,000 database tables and 140 applications. Mapping it had been a multi-month exercise. The FDE team built agents that mapped it in weeks, surfacing the dependencies that Rio Tinto's teams could validate and build against.
“The agents gave us a trustworthy picture of reality in weeks, not months. For the first time, we could see the full shape of what we're modernising.”
- Chris Ware, GM Digital, Rio Tinto Iron Ore
Tapestry
Contract to production in 2.5 months, on Amazon Bedrock AgentCore.
“We are excited about the possibilities agentic commerce can bring to our customers. AWS brought the recipe, but together we built the customization our consumers needed.”
- Yang Lu, Chief Information and Digital Officer, Tapestry
How AWS FDE works
Works runs in 45-day sprints, and every sprint follows the same method, the AI-Driven Development Lifecycle (AI-DLC), with a quality gate at each phase.
1. Before the sprint
Discovery agents map how your business works and draft a living spec in days, with the outcome, the approach, and a fixed sprint price. While you buy, readiness agents assess data, access, and systems and fix gaps.
2. Build and test
From day one, AWS FDEs, your engineers, and agent crews build from proven accelerators in short loops, with real feedback shaping what comes next.
3. Deploy
The agentic system is adapted to your environment, with monitoring, evals, and cost tracking, and one view of every agent for your leaders.
4. Extend
As AWS FDE steps back, your team keeps building. Your engineers grow from observer to operator, and the living context, controls, and code carry into the next use case.
Get started with AWS FDE
FAQs
General
Open allFDE is for organizations committed to AI transformation and ready to move quickly, whether you’re taking your first high-value use case to production or need more speed or scale. The best fit is a priority business outcome with a measurable result, an executive who owns it, data your team can access, and engineers ready to build alongside the FDE team. The problem and your readiness matter more than your industry or size. AWS FDE is available to commercial and public sector customers in 15 countries: U.S., Mexico, Brazil, Canada, Australia, Singapore, India, South Korea, Japan, UK, Germany, Netherlands, France, Sweden, and Saudi Arabia.
- Agentic foundation: build, run, and manage agents across business units on one secure base, grounded in a living model of how your business works.
- Workflow automation: agentifying processes such as claims, document extraction, case triage, loan processing, KYC, and permits.
- AI software factories: AI-driven development, testing, and code modernization at scale, to make existing applications and processes agentic.
- Customer service agents: voice and shopping assistants, plus issue resolution, that transform customer experience.
- Security and vulnerability management: every agent gets its own identity, permissions, and audit trail, and continuous posture management finds and fixes vulnerabilities as AI makes attacks faster.
- Model and inference optimization: tokenomics, routing, and model customization with your own data, tuned to each production workload.
FDE works across industries, including financial services and insurance, healthcare and life sciences, government, sports, media, entertainment and gaming, airlines, retail, automotive, manufacturing, and energy.
Neither. FDE builds with you. Your engineers build with the FDE team from the first sprint, and your team runs and extends the system after the FDE team steps back.
Talk to your AWS account team about the business problems you want to solve.
Delivery and outcomes
Open allConsulting firms and system integrators typically deliver a defined scope against a statement of work. AWS FDE is built to leave something lasting in your AWS account: a production agentic stack, a living context of your business, and a team that can run and extend it. Many consulting firms and systems integrators are AWS Partners, and they extend what FDE builds as you scale.
FDE is sold in 45-day sprints under the FDE Terms, with no separate contract or statement of work to negotiate. Each sprint has a set price based on the team’s location, so you know the cost before you commit. Larger use cases may span multiple sprints. It’s billed through your AWS account, starting when the sprint starts.
You keep using what was built, the agents and workflows in production, and the system it was built in: the living context of how your business works, the workbench, the runtime with guardrails and evaluation built in, the AI control tower to secure, operate, optimize, and govern your agents, and the AI-DLC way of working your engineers learned. It all runs in your AWS account on managed AWS services that stay current.
- An executive sponsor who owns the outcome and clears blockers.
- Access to your environment and data within the first weeks. Synthetic data can start the build while higher-privilege access clears.
- Engineers who build alongside the FDE team from the first sprint.
Your team runs and operates the system. For added support, AWS offers an optional AWS Support engineer who works on production readiness alongside the build and becomes your technical owner at go-live.
Engineering teams
Open allAn FDE team is led by an AWS product lead and tech lead, working with a lead from your organization. Around them are FDEs, scientists, and specialists in areas such as security, design, model customization, and your industry as needed. Many of our engineers come from AWS Service teams, including those that built Amazon Bedrock AgentCore, Kiro, AWS Quick, AWS Transform, Amazon Connect, AWS Security Hub, and AWS Context. They run AI in production at AWS scale, led AWS’s own AI transformation by rebuilding workflows like quote-to-cash and go-to-market, and work side by side with your engineers and crews of AI agents.
Security and governance
Open allEverything runs in your environment, under your identity, network, and security controls, and your data stays there. Guardrails, evals, audit trail, and cost controls are built in from day one, and every agent has its own identity and permissions.
During the engagement, the FDE team threat models the design, runs security scanning in your CI/CD pipeline, and tests before deployment. You approve at defined gates. Before handoff, your team shows it can operate and secure the system, and AWS security and DevOps agents keep scanning and monitoring after that.
You own what we build for you. Anything net-new created during the work is yours, along with your code, data, and documents.
AWS services and programs
Open allFDE and AWS Professional Services work together. FDE takes a business problem, including one you're still defining, to agentic AI running in production, built with your engineers against a measurable outcome. Professional Services delivers defined work against a statement of work, such as a migration. FDE hands statement-of-work projects to Professional Services or AWS Partners, and they extend what FDE built as you scale.
Did you find what you were looking for today?
Let us know so we can improve the quality of the content on our pages