AI NewsGo-to-marketAnnouncement
AWS Forward Deployed Engineering expands to 15 countries with 45-day sprints
AWS now sells embedded engineering in fixed-price 45-day sprints across 15 countries.

Why it mattersBuyers can now compare an embedded AWS team against a direct hire on price and time. That makes forward deployed engineering easier to buy.
Buying an embedded engineering team used to mean a bespoke consulting contract. AWS now sells one in 45-day blocks. The company said its Forward Deployed Engineering service is now live in 15 countries. The service places AWS engineers inside a customer's team. The job is to move AI out of prototypes and into production.
The 15 countries are the U.S., Mexico, Brazil and Canada. The list adds Australia, Singapore, India, South Korea and Japan. It then adds the UK, Germany, the Netherlands, France, Sweden and Saudi Arabia. Both commercial and public sector buyers can use it.
AWS launched FDE earlier this year. The new change is the shape of the service. Each job is a 45-day sprint at a set price. The sprint is billed through the buyer's AWS account. Dr. Asa Kalavade wrote the post. He said the sprints come with pre-built pieces for common jobs. The pieces cover reading documents and reading customer feedback.
What the service does
AWS describes FDE as three layers: a foundation that scales, workflows that run on their own, and governance. The AWS engineers sit with the customer's team through all three. The sprint fits teams that want AI live inside a quarter. It does not suit teams that have been running proofs of concept for a year.
Named customers and their numbers
AWS named four buyers. Marsh, the insurance broker, saw "efficiency gains in the targeted areas" of as much as 70 percent. AWS said some process steps moved "from hours to seconds". AWS did not name the function or the baseline.
Samsung Life Insurance built a governed agent platform in 45 days, per AWS. The platform has "a single control plane and compliance built in". It ran across four affiliates.
Rio Tinto Iron Ore had FDE engineers map a legacy Oracle system in weeks. Rio Tinto's own teams then used that map to build on top.
One unnamed financial services customer saw throughput rise 53 percent, per AWS. Code reviews got 56 percent faster. AWS gave no further detail on the measurement.
Every figure above is AWS's own. The post does not say how the baselines were set. Treat them as the vendor's own claims for buyers.
The buyer who finds this useful
A company that has paid for AI pilots and has nothing live is the audience. The sprint is small enough that a buyer can run one and judge the output. The fixed price is the real change. FDE's older shape started with a scoping phase and a bespoke statement of work. That is hard to compare against a direct hire.
A team that wants to test the embedded model can now ask AWS for a sprint in writing. A team that already runs AI in production is unlikely to need it. Compare the sprint price against six weeks of two full-time engineers before committing.
Source
- Getting AI out of the demo and into businesses globally, AWS Machine Learning Blog, Dr. Asa Kalavade.
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.

