🦄 Making great presentations more accessible.
This project aims to enhances multilingual accessibility and discoverability while maintaining the integrity of original content. Detailed transcriptions and keyframes preserve the nuances and technical insights that make each session compelling.
Overview
📖 AWS re:Invent 2025 - 4x faster workload modernization with agentic AI (MAM349)
In this video, AWS Senior Solutions Architects Lakshmi Dasari and Vimala Pydi introduce AWS Transform, an agentic AI service that accelerates cloud migration and modernization by 4X. They highlight that 70% of workloads still run on traditional data centers and typical migrations take 12-18 months. AWS Transform uses specialized agents trained on 19 years of AWS expertise to automate discovery, planning, and migration phases. The demo showcases VMware migration with automated wave planning, network topology transformation, and VPC deployment. Vector Limited achieved 20-34% faster migration and 35% TCO savings over 5 years. The session also covers the AWS Well-Architected Framework with 57 questions across six pillars for ongoing optimization.
; This article is entirely auto-generated while preserving the original presentation content as much as possible. Please note that there may be typos or inaccuracies.
Main Part
The Migration Challenge: From Personal AI to Enterprise-Scale Transformation
Hi, welcome everyone. I hope you had a fantastic Thanksgiving. I certainly had a great one. I'm just back from a road trip to Arizona, and it was a family road trip. I have two young children, ages 9 and 6, and if you've had screaming kids in the back seat, you know what it's like. It usually causes a little bit of anxiety for us. I wanted the trip to be fun for my kids and also informative and educational, something that they would look forward to. So I turned to Claude, our agentic AI, and created an activity book for them that's personalized for their ages and the places we were going to visit. It was a co-creation process with my daughters. This is what they're growing up with, and I'm so excited about this new technology. This is just a small example of how agentic AI is driving personal productivity.
Here at AWS, we are trying to solve much bigger problems with agentic AI. Today we're going to talk about AWS Transform, our first agentic AI service, and how it is solving the problem of migrations and modernizations. I'm Lakshmi Dasari, a Senior Solutions Architect with AWS, and I have Vimala Pydi with me. Hello everyone. My name is Vimala Pydi. I'm a Senior Migration Solutions Architect from the public sector as well. Let's dive straight in.
How many of you have used any generative AI application so far? That's most of you, right? That's also a validation of the numbers you see on the screen. About 80% of enterprises will be deploying some form of generative AI-enabled applications by 2026. If I were to ask the same question two years ago, less than 5% of your hands would have gone up. More than 15% of work decisions will be automated by agentic AI by 2028, and that's up from 0%. We're truly talking about autonomous decisions made by agentic AI, however, through human oversight. The human oversight is the most important part.
Yet there's a lot of technical debt out there. 70% of workloads are still running on traditional data centers and on-premises technology. More than 70% of legacy IT software is written more than 20 years old. What does that mean? There aren't a lot of people who understand the technology. It's hard to maintain, it's brittle, and it's hard to modernize. How do we then move from the current state to a future state?
If you look at a typical cloud adoption journey, you may each be in different phases of the cloud adoption journey. The real value realization happens when the organization makes a deliberate decision to move away and retire the technical debt and adopt cloud-native technology. That's when the real reinvention happens. One of the segments I was supporting earlier in my previous role was digital-native business customers. These are customers that are born in the cloud, constantly innovating and constantly reinventing. Yet it typically takes about 18 plus months to complete a migration.
Migrations are hard, they're complex, and they involve a lot of activities with many teams coming together. If you look at a typical migration methodology, there are typically three different phases: assessment, plan, mobilize, and migrate and modernize. One of the first things that happens in an assessment is creating a business case, the case for change. Why is this important? What is the ROI? What is the benefit for the business? Once you've done that and got the executive buy-in, you do a deeper discovery into your IT state, where the applications are in the lifecycle, and then determine the migration strategy from there on.
Many of them may be just rehosted from the on-premises environment into AWS, from traditional VMs into EC2s, and some may need retirement, re-platforming, or rearchitecting. Once you've determined the migration strategy, you move on to building a migration plan. This is where you determine which servers can be moved in waves based on how dependent they are or what environment they run in. Then you build your landing zone architecture in AWS, which creates a strong foundation for the subsequent migrations to come.
Now imagine doing this process for thousands of servers. This is where agentic AI can rapidly accelerate that process with the help of LLMs and graph neural networks. We can process millions of lines of code, extract business logic, map application dependencies, the network topology, and the infrastructure that they're dependent on, and be able to create the target infrastructure artifacts that are needed for a successful migration.
AWS Transform: Accelerating Migration with Agentic AI and Specialized Agents
So this is where AWS Transform comes into play. With that, I pass it to Vimilla. I would want all of you to pause and think about the data point Lakshmi shared with us. Typical migration takes from 12 to 18 months, about a year and a half. Today we are going to talk about AWS Transform, an agentic first service, which would accelerate your migration and modernization by 4X times.
It has specialized agents built for VMware, mainframes, Windows, any stack, and it has any-to-any custom transformation as well. What differentiates AWS Transform from its previous version is that it is built on migration methodology. It has scope to do end-to-end migration, as is mobilize, migrate, and modernize that Lakshmi spoke about. It has specialized agents which are trained on thousands of migrations over 19 years of AWS expertise.
It has a unified web experience where users can log in and get a preview of their migration at any point in time. They can use it with enriched chat functionalities. They can also use it to collaborate between their teams. It also has the right balance between autonomy and human-in-the-loop. I'm going to show a demo which will look at how we can use the end-to-end migration within VMware.
This is how the experience would look like. Once you get into AWS Transform, you have a workspace which is a container for your transformation. The first step is to create a job. This is where you define your migration scope. It can be a mainframe migration, any of the specialized migrations, discovery of a specific workload. But for this demo purpose, we are going to pick migration. I am going to pick VMware migration.
One of the key features that we have with Transform is it has different customized steps that you can choose for which aligns with your organization. You can either choose end-to-end migration, you can just do discovery or just network migration. Today, let's go ahead and do end-to-end migration, and that's it. It gave us a roadmap. This is the roadmap for your migration, for your transformation. It is aligned to the migration best practices.
Phase one is discovery and planning, phase two is infrastructure setup, which is mobilize, and the third one is migration and modernization phase. You also have the chat functionality here. If you want to customize this, you can provide your input and it would customize your roadmap as well. I'm going to start the migration, create the job, and once I create it, it gives me a job plan. Again on the left, these are the steps that AWS Transform tool is going to run for us. We would be able to automate it or we are going to look at how we can provide feedback as we are moving through these steps.
For today's demo purpose, I'm going to go over some key features. If you want to look at the complete demo, please stop by our kiosk at Migration and Modernization, and we'll go through all the steps here. As you can see, the first step here is importing inventory. That's the discovery phase. You can either use our discovery agent to know your source environment, or you can import any of the RVTools exports, CMDB, any of your custom inventory that you have. You can definitely put that in through this tool.
Let's take a look at migration plan. Once you import the inventory, we are going to look at migration plan.
This is step one. It gives us a preview of the input inventory that we loaded. If you look here, it has about 48 total servers. It provides us the breakdown of network connectivity, and it also created output files from here. If you want to customize your migration plan, this would help us automate and create your waves. But if you would like to customize that, you can use the chat functionality to customize it.
Here I'm going to use the scope of VMware migration. I ran a report, so I do have the scope, which is for specific environments, and it's pulling up only VMware. Another key point is that at any task level, if you want to provide instructions, you use E followed by the task name. So here I'm saying E build migration plan, and I'm asking you to create a very custom wave plan for me based on the VM information that I'm giving.
Let's look at how it looks like when the waves are created. Behind the scenes, there are a few steps which you can give feedback on how the wave planning will be done. You can group the servers or applications and provide that feedback, but this is the final output. It's giving us four wave plans recommending us to do it sequentially. If you're happy with it, you can complete it and move forward to the next step.
After the migration plan is created, since it's end-to-end migration, you would be able to connect to your target accounts. Transform gives you a transform connector. The tool will walk you through the information on how to input the Transform connector into your target accounts. Once you do that, the next one is migrating network. Network migration is one of the key elements where it's simplifying complex source network topology into aligned target networks in AWS.
Here I'm going to showcase how it is going to take in your input. We can put in RVTools, NSX, and we have latest support for Cisco, Palo Alto, and Fortinet files as well. We are going to upload a specific network topology into it. The tool will take that input. If you want to change, there are two ways that you can look at your network topology when it comes into AWS. You can either use a hub and spoke method or you can do isolated VPCs.
If you choose isolated VPCs, you would be able to update your CIDRs too. This is how you can go ahead and edit your CIDRs. For demo purposes, we are going to put in a new CIDR. And save. We'll go step into the last phase here. Once you're happy with your network, the tool allows you to create your approvals as well. This is where you can really get into the RBAC and have different teams work on different migration plans.
Here we are using the demo user as the approval, so we'll go ahead and approve it. Once approved, this would automate the VPC deployment into your target applications. The last part of it is executing your migration plan. Under the hood, we are using MGN connector for executing your migration plan. It goes in and gives you details on how you can install your application agents into your source environment, and it goes through the process of getting your replication, continuous replication, testing your environment, and then doing a cutover.
Those three steps are how it works under the hood on executing your migration plan. This is a preview of how AWS Transform works. If you want to look at the complete demo, please stop by at our kiosk after the session. I and Lakshmi will be hanging out there too. If you have any questions, please feel free to come and ask. Lakshmi, over to you.
Yeah, great. Thank you, Vimla. I quickly wanted to talk about the customer success story here. Vector Limited is a New Zealand-based energy company and they were able to migrate from VMware Cloud on AWS into native AWS architecture using AWS Transform for VMware.
With the help of a partner Slalom, they were able to achieve 20-34% faster migration time and savings of 35% for the next 5 years in TCO. The biggest help they had that contributed to the faster migration is the wave planning, which is one of the tedious steps in any migration. Understanding your application dependencies, grouping them into waves, speaking with the application teams, and then doing the migrations themselves are all critical components of this process.
Beyond Migration: Maintaining Architecture Excellence with AWS Well-Architected Framework
Now that we have gone through this cloud journey together, let's say you're migrated some EC2 instances and some services from on-premises into EC2, and then you are operating your workloads in AWS. How do you then keep up with the technical debt that creeps into your technology? This is where the AWS Well-Architected Framework and Well-Architected Tool come into play. This is a tool that's available in the AWS console now as a self-service tool with a guided experience.
You define a workload, you go through the guided experience of answering questions, and then you find remediations for the risks that it identifies. This is based on our collective best practices over all of the evaluations that solutions architects have done in the last several years for our customers. How do you say you're well architected? It's well architected if it's well architected across these six pillars. More often than not, we find that most applications perform well in some pillars, whereas they don't perform well in some of the other areas. There are about 57 questions across these pillars, and you go through that guided experience and then you prioritize which pillars you want to work on.
Maybe cost optimization is a major pain point for you as an organization. Maybe reliability is a major need, as business continuity and disaster recovery plans are something that are pretty much a must for every organization. What does the architecture improvement framework workflow look like? First, you identify your workload. You bring the people into the process, then you go through the guided experience identifying the pillars that you want to prioritize. You answer the questions for each of those pillars, and at the end of it, you get a report that identifies high-risk findings as well as medium-risk findings.
The high-risk findings are those that need to be addressed and they have significant impact to the business. Once you've understood the risks, you can find prescriptive solutions in the framework itself, and you can prioritize which improvements you want to work on and accelerate the implementation of those findings. This is where tying back to the agentic AI solutions comes in. This is where you can use AWS Transform for VMware, Windows, and custom mainframe solutions to really accelerate that journey of modernization further.
With that, we conclude the session. We really hope you enjoyed and learned something in today's session. There is a migration and modernization kiosk right in this AWS village, and we'll be there as well. If you have any questions, please come and talk to us. If you haven't done an EBA, which is an experience-based acceleration program that we have, within a week or so, multiple teams can come together, take a small workload, go through the whole process, and it's the best way of learning. Please do talk with your account managers to schedule an EBA. Thank you, and feel free to connect with us on LinkedIn as well. Please, please, please do take the survey. It's very important for us, and it gives us very qualitative feedback on how we're doing. Thank you so much.
; This article is entirely auto-generated using Amazon Bedrock.
Top comments (0)