🦄 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 - Monetizing agentic AI solutions: A Partner’s guide for AI agents & tools-PEX112

In this video, Seenu Brahmarouthu and Leno Piperi guide partners on monetizing agentic AI solutions through AWS. They explain the evolution from Generative AI assistants to autonomous multi-agent systems, showcasing Druva’s success in resolving 63% of security issues autonomously using Amazon Bedrock Agent. The session covers four product strategies from AWS’s "Rethinking SaaS in the Agentic Era" white paper: extending, building, embedding, and expanding. They introduce the AI Agents and Tools Marketplace launched in July, which offers three categories: AI agents, tools including MCP servers, and consulting services. Key benefits include flexible pricing models (usage-based and contract pricing), multiple deployment methods (API-based and container-based), and tight integration with Amazon Bedrock AgentCore. The demonstration shows how customers can subscribe and deploy agents with just a few clicks, with automatic usage tracking and payment processing through AWS Marketplace.

; 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 Evolution of Agentic AI and Strategic Frameworks for Partner Monetization

Good evening everybody. My name is Seenu Brahmarouthu, and together along with my colleague Leno Piperi, we’re excited to talk to you about monetizing agentic AI solutions. This is a session designed specifically for partners to provide you some guidance on your journey on agentic AI and how to monetize and drive revenue.

It’s fascinating in terms of Generative AI how we got here and the different waves of AI and how we got to agentic AI. If you think about it, there are a couple of different waves that we’ve gone through. A couple years ago with Generative AI assistants such as Claude and ChatGPT, we were able to automate some of our day-to-day tasks. Then as we advanced into Generative AI agents, which are powered by much more powerful advanced models with reasoning capabilities, we were able to automate some of the more complex tasks. Now as we move into the agentic space where everything is much more fully autonomous, we’re seeing multi-agent type of systems.

The definition from AWS in terms of what an AI agent is is an autonomous system that can plan, reason, and take action on behalf of a human. As you think about where we’re going into the future, we believe there will be billions of agents out there, and this is a great opportunity for partners to take advantage of this wave and show you some tips and tricks in terms of how you can work with AWS.

You may have heard about all the different services that we’ve launched, but just a quick recap. All the way from applications to development tools to infrastructure that we offer, we have applications such as Amazon Q, Amazon QuickSight, as well as AWS Marketplace, which we’ll double click on in this session today. We also have development tools such as Bedrock to fine-tune your models, as well as Bedrock AgentCore, which is the building blocks for building your agents specifically. Then on the infrastructure layer, we have SageMaker, which allows you to build and deploy your own models, and finally compute with Trainium chips that we offer, as well as the GPUs that we have partnerships with NVIDIA.

Just taking a step back and looking at one example of an ISV partner, since this is a partner session and partner guidance, Druva is one of our ISV partners in the cloud data security space, and they had a challenge with their team in terms of being able to resolve all of these security issues for their customers. The team was getting really overwhelmed with the complex number of security workflows that they had to resolve, so they built a solution called Dru AI, which is embedded within their SaaS application. It’s a multi-agent system where their solution called Dru AI is built on Amazon Bedrock Agent, and they offer a multi-system type of architecture where there are data agents, insight agents, and action agents that give the ability for customers to resolve issues. The results speak for themselves. They’re able to resolve 63% of the customer security issues autonomously. So just one example of a partner that has adopted agentic AI as part of their SaaS solution, and they’re seeing tremendous results.

Now, feel free to take a QR code to pull up the link here, but I wanted to just share, as a partner going through the journey of agentic AI, what are some of the product and go-to-market strategies for you to consider in your journey to bring agentic capabilities within your own offerings. AWS published a white paper a couple months ago called Rethinking SaaS in the Agentic Era, and it really boils down to four different strategies that you can think about. The first one, and both of those strategies are dependent based on product and go-to-market. With product, you’re thinking about new features, either new features or completely new products, and from a go-to-market perspective, it’s either your existing customer base or new customer base that you’re going after. So let me go one by one just to share at least from a high level.

The first strategy is around extending. This is around new users within your existing customer base.

You’re expanding the scope of your existing products with new features and new agentic capabilities for new users within your existing customer base. In this approach, you can possibly have an upsell as a strategy and charge a little bit more as you’re extending and bringing more agentic capabilities. If you look on the right-hand side, on the far right on top, it’s building new products and targeting new customers with new personas and new markets. There’s definitely some risk involved here because you’re building something forward-thinking, but this is where some of the opportunity is in terms of new revenue. This could be a sales agent as a service or a marketing agent as a service as we move into this agentic space. You’ll start to see a lot of these types of offerings from partners.

Now on the bottom left, embed. This is for new features for existing customer base. You’re embedding AI features, and in this quadrant you’re probably not charging anything extra, but this could be a really good retention strategy if you’re in a competitive market and you want to make sure you retain your customers with more value. Embedding and bringing those agentic capabilities could be one of the approaches that you can take. Finally, in terms of expanding, this is new products with existing customer base. You can have some complementary solutions along with your existing products where you’re still targeting your existing customer base, but it could be an additional revenue stream for you.

So just a quick framework to kind of think about how to bring agentic capabilities within your offering, within your solution. Feel free to take a QR code of that, and it goes into much more detail on the white paper that we have.

AWS Marketplace AI Agents and Tools: Flexible Pricing and Deployment Options

Now that’s great. Once you figure out what your agentic capabilities are, how do you go to market with AWS? Earlier this year we launched the AI Agents and Tools Marketplace in July, and this is our one-stop shop for customers to find all of our agentic offerings from our partners. When partners come to AWS Marketplace, they expect to see agentic type of offerings, and we’ve categorized our AI Agents and Tools section within AWS Marketplace so they can find them in three different categories. The first category is AI agents. This is where you can find some pre-built agents or agents that are embedded into software applications, or even agent development platforms that allow you to build your own agents. So that’s one workstream of potential solutions that customers will find.

Customers are also looking, as they’re building their agentic workflows, for tools. We have a lot of partners already listing their MCP servers on AWS Marketplace, so that’s another route that’s possible. Finally, for consulting partners, we have many consulting partners that provide AI strategy consulting as well as implementing that. All of those offerings get bubbled up to our customers when they come to AWS Marketplace and look in the AI Agents and Tools Marketplace.

Now there’s a lot of benefits associated with the AI Agents and Tools that are in AWS Marketplace. I’d like to have my colleague Leno come up here and talk a little bit about that. Thank you, Seenu. So this is a partner session. I want to get a quick poll. Who here has a listing in Marketplace today? Perfect, so a good number of you. Now for those that don’t, allow me to convince you. AWS Marketplace is a curated digital catalog that makes it really easy for AWS customers to find your products, subscribe and purchase your products, and then deploy them in the location that they need them.

Partners that lean into AWS Marketplace benefit from deal acceleration from the procurement efficiency that they get from AWS Marketplace. Also, when you sell on AWS Marketplace, you’re aligning with the broader AWS strategic goals, so you get to grow with us. Lastly, partners that are in AWS Marketplace get to leverage the vast AWS co-sell network that has global reach. And now AWS Marketplace is also the best place to list your agents and tools. Let’s understand why.

Flexibility is going to be a theme that I bring up a few times here. Marketplace gives you the flexibility to price your agents and tools how your customers want to purchase them. We offer usage-based pricing, and this is great for all customers. Everyone wants usage-based pricing, but it’s particularly important for your smaller customers or one-off developers that want to kick the tires on an agent or tool and pay for what they consume before they commit to a broader contract. The way this works in the Marketplace is that your agent will track that usage and report it to Marketplace, and then Marketplace will automatically bill the customer and disperse payment directly into your bank account.

We also offer contract pricing, and with contract pricing, this is an obvious favorite for enterprise customers. You’re able to transact upfront payments and you can have broad durations in that contract, and we make it really easy to manage your licenses with that contract by giving you access to several tools. It’s easy for you to check entitlements on contracts that your customers purchase. Both pricing models are supported by AWS Marketplace private offers. A private offer in Marketplace is an agreement between you and a customer so you can negotiate terms that are favorable to both parties and then transact that directly in AWS Marketplace. When you transact through private offers in Marketplace, you can set up flexible payments to make it really convenient for you and your customers to disperse payments on a schedule that makes sense for both of you.

Continuing on the flexibility, AWS Marketplace makes it easy for you to sell your agents and tools with a deployment method that makes sense for your customers. We offer API-based agents where you’re hosting the agent in your own AWS account and you’re giving access to the customer via an endpoint. They can then integrate into their own systems. We also offer container-based agents and tools. Here you’re defining your agent as a Docker image that is hosted in AWS Marketplace, and then when the customer purchases that product, they’re given access to that image and they can deploy it in the container orchestration service of their choice.

We also support professional services and additional software offerings as well. You can offer your design and implementation services to complement your agents and tools, and you can always sell SaaS that has agentic solutions embedded in it as well. Now hopefully this is enough to pique your interest in selling your agents and tools in AWS Marketplace, but the reason why I think it’s the best place to sell your agents and tools is because AWS Marketplace is tightly integrated with Amazon Bedrock AgentCore.

Integration with Amazon Bedrock AgentCore and Customer Subscription Workflow

Hopefully by now you’ve gotten a lot of great information on Amazon Bedrock AgentCore. Just a quick refresher: AgentCore is a platform that allows you to easily develop, maintain, and operate agents and tools directly on AWS, and we think it’s how customers are going to want to develop the future of their agentic systems. Those same customers want access to partner agents and tools so that they can make their development both easier and more productive. For the partners, AgentCore offers a ton of flexibility. We support multiple frameworks, all the leading frameworks, all the foundation models available on Bedrock, and all the leading protocols like AAA and MCP. Your solutions might already be AgentCore compatible.

When you list your API listing in Marketplace, you make it that much easier for customers to connect it to their solutions with just a few lines of code, and the same goes for the container-based offering. Let’s see how this works in practice. It starts with you listing a product in AWS Marketplace. In this run-through, I’m going to show the flow for both container-based and API-based listings. Then you have a buyer that subscribes to your offerings. For the container path, once a buyer subscribes through Marketplace, they’re guided through a closely controlled workflow that eventually results in them deploying an agent directly in AgentCore Runtime. AgentCore Runtime is the serverless hosting environment for AgentCore.

For the API-based deployment method, that same customer is again guided through Marketplace, provided with an endpoint and the necessary keys to use that endpoint, and they get access to your agent which is running in your own seller account.

Now the buyer will take these agents, integrate them into their applications to enhance their own applications, make them agentic as they consume. Each agent is reporting that usage to AWS Marketplace. And then Marketplace is billing the buyer and then dispersing payment directly into your seller account. So that’s the whole flow.

Now, I’d like to show you what this looks like from the actual customer experience just to show visually how easy it is for a customer to subscribe to and use an agent from AWS Marketplace. This starts, so this is an example listing of a container-based agent in AWS Marketplace. And here you’re seeing all the various things that you can show on your detail page. You can have videos, everything.

With one click, you’re sent into a subscription experience here. A customer can view the various usage costs associated with this product, terms, conditions, and they have a few configurations here. With one click they can subscribe. In a couple of minutes, the subscription will go through and they click a button to launch that agent.

Now they select Bedrock AgentCore as a deployment method. They are given some instructions for what they will need to do to make the most use out of that agent, but they have a one-click button there to actually deploy that agent directly in AgentCore. That’s great. And now you have an endpoint that’s ready to use and integrate directly into your agentic system solutions, so that could be your product right there.

To close off, I want to give you access to some resources that will help guide you through your next steps. If you’re interested in listing in Marketplace, in general, you can find everything over in our seller guide, and it’ll point you to all the resources for all the various products that you can have in Marketplace. Also, please do take the time to explore more and learn about Bedrock AgentCore if you haven’t already. And lastly, I’ll invite you to learn more about joining the AWS Generative AI Competency. It’s a new competency for the partner network, and we think you’re going to be really excited about it. Thank you, everyone.

---

; This article is entirely auto-generated using Amazon Bedrock.