AWS Summit NYC 2026 | Jose Kunnackal, AWS & Russell Jukes, DXC Technology

Clips
More from AWS Summit NYC 2026

Jose Kunnackal

Director, Amazon Quick

AWS

Russell Jukes

Chief Digital Information Officer

DXC Technology

play_circle_outline Deploying Agentic OS at Scale: Amazon and DXC on Productivity, AI Fluency, Guardrails, Token Economics, and Partners

play_circle_outline Personal agents versus professional agents need different context types.

play_circle_outline Faster adoption in repetitive, binary workflows; high-discretion tasks slower to automate.

Info

Jose Kunnackal, AWS & Russell Jukes, DXC Technology

Jose Kunnackal

Director, Amazon Quick AWS

Russell Jukes

Chief Digital Information Officer DXC Technology

search

Gemma Allen

>> Welcome back to theCUBE coverage here live on the ground at the Javits Center in New York City. This is AWS Summit 2026 and we are bringing you all of the action. Joining me now are two folks key to that story, Jose Kunnackal, director of Amazon Quick, and Russell Jukes, partner at DXC. Welcome, folks. Let's talk about some of the learnings from that journey because we hear so much about AI for enterprise. The opportunities are endless, that it's going to reshape everything we've ever known. We also know that in reality, some folks say that it's not as easily done as we're made to believe, right? There are certain challenges and I guess requirements needed to bring this to life. Talk to us about some of the learnings that I guess you guys have seen, especially in this relationship and partnership you've built together.

Jose Kunnackal

>> Oh, I can definitely talk to several. I think we can talk about a few and I'll hand it over to Russell, because I think some of these have been how really you can make change happen in the enterprise. When you think about AI, it's easy to think of it as a chatbot or something that people jump to get an answer. But really the question is, how does it fundamentally transform the way you work? So with Amazon Quick, we've built something that is proactive and personal. And so the AI is not about the system that it's in, it's all about you. So helping you write a better document, look across all of the information you have in order to get to the insights that you want, and then get to the actions that you need, whether it's as simple as sending an email or kicking off an entire workflow, but at the same time being connected to that context. All of that is important to make AI useful because otherwise you're never going to use it. But once it becomes useful, the question is, how do you roll it out across the company? And that's where we've had many learnings together. And as we rolled it out at DXC, there are about 115,000 users using it. At Amazon, we've got many hundreds of thousands of users using it. And so some of the learnings there is what I'll let Russell jump into.

Russell Jukes

>> Yeah. This is a super complicated topic, right? Not because it's hard to understand, but because there's many elements to it. I think the first learning I had was do not deploy AI like it's a SaaS app. Don't do that, because I tried that and it didn't work. I also used to measure the deployment of technology and then I used to measure the adoption of technology, and that doesn't work either because AI fluency of the people using the technology is more important than the adoption and the deployment of it. So getting into measuring, and everyone asked me, "How do you measure success here?" I'm measuring the fluency of our people using the technology. And that's a different set of measures that we would've used before. There's a whole set of things around guardrails and how do you manage the token economics of the world and all these kind of buzzwords. But at the end of the day, picking a good partner that understands how it works, that's done it before really makes the biggest difference.

Jose Kunnackal

>> I'll just add to that, that rolling out AI is just part of the story. So to what Russell was saying, AI fluency of the users, like helping them be comfortable with it and really rolling it out not just at a user level, but also figuring out how can a team use it, setting some of the guardrails, some of the pieces that really make it work in a team and an enterprise. Because ultimately when you roll out something like Quick, you are rolling out an Agentic OS across the company. Really it becomes the way through which you can deliver other agents that you build, that you custom-built for your organization. And of course you make sure that every user has access to it.

Gemma Allen

>> That's a very interesting observation, don't roll it out like it's a SaaS app. I actually think that there's so much to unpack there. But one thing we know is that AI is context hungry, right? For it to be effective, it needs to own and understand and continue to evolve the context layer.

Russell Jukes

>> Yes.

Gemma Allen

>> What are the best practices in terms of how you capture context within an enterprise?

Russell Jukes

>> I made a division very early on that we would have two types of agents in our company. One, the personal agents that you and Jose could create for yourself to drive personal productivity; but then there would be professional agents, ones that me as an IT professional will create for the entire enterprise, and that will drive enterprise value. And so they need different types of context. You need context all about you and the job you do and the role, and professional agents need much more broader end-to-end workflow context. And so if you think about it in those two things, you'll actually start your journey on a good path.

Gemma Allen

>> When we think about Amazon Quick broadly and we think about the opportunities it brings to organizations to work more efficiently, more fluidly, and also to connect parts of the businesses that perhaps were quite siloed up until very recently, we all know the long-term game of siloed applications and workflows. How do you see that shift happening? Do you think it's happening at a practice by practice level? At a broad kind of overarching enterprise level? What really needs to happen for enterprises to understand, "Okay, there is a huge opportunity here. We need to think differently."

Jose Kunnackal

>> I think it's an interesting question, just connecting it back to what you asked earlier, because context actually resides in multiple silos today. The information that you need is split across SharePoint or Google Drive, across your business metrics, across files that people have. And bringing all of that together is a challenge by itself. So one of the things that Quick does really well is it connects to your sources. So it doesn't matter if you use the Microsoft ecosystem, the Google ecosystem, Salesforce, Quick connects to those sources and it enforces the rules that they have. So when you're a user, you log in, you get to see what you see, and then it brings that context together with what's on the internet. It can do a web search, it can do research. Then it brings that together, and again, files on your desktop. One of the most important things we found, again, working together is there's files on your desktop that might be super critical to you, but those are not connected to what's in the cloud. This is what you just wrote down. When you connect all of this, you're bringing that unified context together and so you're really breaking down those silos. So it doesn't matter where the information is, you bringing it together and now you're able to create a better PowerPoint or craft a better project plan. Or when you go for a sales pitch, you know what was discussed in Teams or Slack the last hour. You also know all the context that was in your email from the last five meetings and of course the notes that you had. Bringing those pieces together and out of those silos is one of the things that Quick is doing really well. And especially with the new desktop app that we have, people are able to get to that in a singular place on their laptop and that's really been game changing.

Gemma Allen

>> Talk about the types of opportunities and I guess fast adoption you're seeing at an industry level. Obviously DXC has worked across a lot of different industries, as does Amazon and Amazon Quick. Are there particular verticals that have been quite accelerated, accelerated adopters of Amazon Quick broadly? Or I guess a low hanging fruit per se that we may not know of? What are you actually seeing? Who is fast and who is slow to adopt this?

Russell Jukes

>> Well, there are some things about speed and AI. Jose and I always joke that the number of agents that are deployed in an enterprise is not your measure of success. We thought it was, and then somebody, "How many tokens you use? Is that your measure of success?" And then we're like, "No, that's not your measure of success either." So it's not going to be vertically integrated. What we're really seeing is if you can create a context graph, let's just call it a graph, of everything that you do in the work at work, every decision that you make, every decision waiting on you, every approval you make, every approval waiting on you, every document and word you write, every email you send, everything you do at work together, the value of that of not being able to get to it today, the value of being able to get to that is where enterprises will see the true kind of exponential framework that they can get to. So legal firms will go down one path; that's kind of a whole different thing. Insurance companies will go down another. But at the end of the day, everybody's looking for that full context aware graph of everything that happens in a daily life. Now, the people moving quicker are the people in repetitive task actions where there's not so much, let's say, human... Let's say discretion in the decision. If it's binary, if we do this, it turns into... So once the binary actions come out, that's where people are able to move super quickly. Once we get into lots of human discretion, Quick still enables you to do that, but that's where you move a little slower. And I think rightfully so. I think AI isn't here to replace everybody and get rid of everybody's job. It's here as a superpower to kind of enable you to be better.

Jose Kunnackal

>> Actually, if I add onto that, one thing, the part where you need more discretion, Quick also provides this thing called an activity feed. So we launched a number of items today with regards to that. But what the activity feed does, it goes across all these things that are happening in your email. So you could have a long conversation with like 10 different threads on something. Pulling that together in order to think about, "What is the point of view? What is your decision needed? What is the next step?" And surfacing the most important ones, especially when you have to make approvals and decisions. The activity feed ends up being a singular hub in Quick for most users today that use us. And what we're seeing is that change, you ask a question about change, is it fast? Is it slow? Once people get the hang of it, that AI fluency part that Russell was talking about, it becomes a second nature to them to do something in Quick. Now the part that's interesting is also that as a product, we've also made it extremely self-aware. Quick knows what it can do and how it can do it. The traditional model would've been you roll something out and then you teach people, "Hey, this is how you click this button, this is what you do." So again, don't go back to the SaaS way of doing things. Do it because AI is different. You've got to think about it differently. People have to understand that. That's the AI fluency part. But once they get it and once they start picking up the tool, adoption is extremely fast. Because you'll find that they find it a new way of working. It becomes a surface where they do a lot of their tasks, from approvals to sending a quick response to an email. All of that's possible from the same surface and that really helps accelerate those deployments.

Gemma Allen

>> Well, it certainly provides a new level of synthesis, right? Which I think we all need and enjoy in our lives in a productivity perspective. So folks, thank you so much for telling the story and thanks for joining us on theCUBE.

Russell Jukes

>> Great. Thank you very much for having us.

Jose Kunnackal

>> Thank you.

Gemma Allen

>> I'm Gemma Allen coming to you from the AWS Summit here at the Javits Center in New York City. We are bringing you all of the coverage on the ground, talking to some of the builders, breakers, and innovators generating this next wave of AI for AWS. Don't go anywhere.

Jose Kunnackal, AWS & Russell Jukes, DXC Technology

search

Gemma Allen

>> Welcome back to theCUBE coverage here live on the ground at the Javits Center in New York City. This is AWS Summit 2026 and we are bringing you all of the action. Joining me now are two folks key to that story, Jose Kunnackal, director of Amazon Quick, and Russell Jukes, partner at DXC. Welcome, folks. Let's talk about some of the learnings from that journey because we hear so much about AI for enterprise. The opportunities are endless, that it's going to reshape everything we've ever known. We also know that in reality, some folks say that it's not as easily done as we're made to believe, right? There are certain challenges and I guess requirements needed to bring this to life. Talk to us about some of the learnings that I guess you guys have seen, especially in this relationship and partnership you've built together.

Jose Kunnackal

>> Oh, I can definitely talk to several. I think we can talk about a few and I'll hand it over to Russell, because I think some of these have been how really you can make change happen in the enterprise. When you think about AI, it's easy to think of it as a chatbot or something that people jump to get an answer. But really the question is, how does it fundamentally transform the way you work? So with Amazon Quick, we've built something that is proactive and personal. And so the AI is not about the system that it's in, it's all about you. So helping you write a better document, look across all of the information you have in order to get to the insights that you want, and then get to the actions that you need, whether it's as simple as sending an email or kicking off an entire workflow, but at the same time being connected to that context. All of that is important to make AI useful because otherwise you're never going to use it. But once it becomes useful, the question is, how do you roll it out across the company? And that's where we've had many learnings together. And as we rolled it out at DXC, there are about 115,000 users using it. At Amazon, we've got many hundreds of thousands of users using it. And so some of the learnings there is what I'll let Russell jump into.

Russell Jukes

>> Yeah. This is a super complicated topic, right? Not because it's hard to understand, but because there's many elements to it. I think the first learning I had was do not deploy AI like it's a SaaS app. Don't do that, because I tried that and it didn't work. I also used to measure the deployment of technology and then I used to measure the adoption of technology, and that doesn't work either because AI fluency of the people using the technology is more important than the adoption and the deployment of it. So getting into measuring, and everyone asked me, "How do you measure success here?" I'm measuring the fluency of our people using the technology. And that's a different set of measures that we would've used before. There's a whole set of things around guardrails and how do you manage the token economics of the world and all these kind of buzzwords. But at the end of the day, picking a good partner that understands how it works, that's done it before really makes the biggest difference.

Jose Kunnackal

>> I'll just add to that, that rolling out AI is just part of the story. So to what Russell was saying, AI fluency of the users, like helping them be comfortable with it and really rolling it out not just at a user level, but also figuring out how can a team use it, setting some of the guardrails, some of the pieces that really make it work in a team and an enterprise. Because ultimately when you roll out something like Quick, you are rolling out an Agentic OS across the company. Really it becomes the way through which you can deliver other agents that you build, that you custom-built for your organization. And of course you make sure that every user has access to it.

Gemma Allen

>> That's a very interesting observation, don't roll it out like it's a SaaS app. I actually think that there's so much to unpack there. But one thing we know is that AI is context hungry, right? For it to be effective, it needs to own and understand and continue to evolve the context layer.

Russell Jukes

>> Yes.

Gemma Allen

>> What are the best practices in terms of how you capture context within an enterprise?

Russell Jukes

>> I made a division very early on that we would have two types of agents in our company. One, the personal agents that you and Jose could create for yourself to drive personal productivity; but then there would be professional agents, ones that me as an IT professional will create for the entire enterprise, and that will drive enterprise value. And so they need different types of context. You need context all about you and the job you do and the role, and professional agents need much more broader end-to-end workflow context. And so if you think about it in those two things, you'll actually start your journey on a good path.

Gemma Allen

>> When we think about Amazon Quick broadly and we think about the opportunities it brings to organizations to work more efficiently, more fluidly, and also to connect parts of the businesses that perhaps were quite siloed up until very recently, we all know the long-term game of siloed applications and workflows. How do you see that shift happening? Do you think it's happening at a practice by practice level? At a broad kind of overarching enterprise level? What really needs to happen for enterprises to understand, "Okay, there is a huge opportunity here. We need to think differently."

Jose Kunnackal

>> I think it's an interesting question, just connecting it back to what you asked earlier, because context actually resides in multiple silos today. The information that you need is split across SharePoint or Google Drive, across your business metrics, across files that people have. And bringing all of that together is a challenge by itself. So one of the things that Quick does really well is it connects to your sources. So it doesn't matter if you use the Microsoft ecosystem, the Google ecosystem, Salesforce, Quick connects to those sources and it enforces the rules that they have. So when you're a user, you log in, you get to see what you see, and then it brings that context together with what's on the internet. It can do a web search, it can do research. Then it brings that together, and again, files on your desktop. One of the most important things we found, again, working together is there's files on your desktop that might be super critical to you, but those are not connected to what's in the cloud. This is what you just wrote down. When you connect all of this, you're bringing that unified context together and so you're really breaking down those silos. So it doesn't matter where the information is, you bringing it together and now you're able to create a better PowerPoint or craft a better project plan. Or when you go for a sales pitch, you know what was discussed in Teams or Slack the last hour. You also know all the context that was in your email from the last five meetings and of course the notes that you had. Bringing those pieces together and out of those silos is one of the things that Quick is doing really well. And especially with the new desktop app that we have, people are able to get to that in a singular place on their laptop and that's really been game changing.

Gemma Allen

>> Talk about the types of opportunities and I guess fast adoption you're seeing at an industry level. Obviously DXC has worked across a lot of different industries, as does Amazon and Amazon Quick. Are there particular verticals that have been quite accelerated, accelerated adopters of Amazon Quick broadly? Or I guess a low hanging fruit per se that we may not know of? What are you actually seeing? Who is fast and who is slow to adopt this?

Russell Jukes

>> Well, there are some things about speed and AI. Jose and I always joke that the number of agents that are deployed in an enterprise is not your measure of success. We thought it was, and then somebody, "How many tokens you use? Is that your measure of success?" And then we're like, "No, that's not your measure of success either." So it's not going to be vertically integrated. What we're really seeing is if you can create a context graph, let's just call it a graph, of everything that you do in the work at work, every decision that you make, every decision waiting on you, every approval you make, every approval waiting on you, every document and word you write, every email you send, everything you do at work together, the value of that of not being able to get to it today, the value of being able to get to that is where enterprises will see the true kind of exponential framework that they can get to. So legal firms will go down one path; that's kind of a whole different thing. Insurance companies will go down another. But at the end of the day, everybody's looking for that full context aware graph of everything that happens in a daily life. Now, the people moving quicker are the people in repetitive task actions where there's not so much, let's say, human... Let's say discretion in the decision. If it's binary, if we do this, it turns into... So once the binary actions come out, that's where people are able to move super quickly. Once we get into lots of human discretion, Quick still enables you to do that, but that's where you move a little slower. And I think rightfully so. I think AI isn't here to replace everybody and get rid of everybody's job. It's here as a superpower to kind of enable you to be better.

Jose Kunnackal

>> Actually, if I add onto that, one thing, the part where you need more discretion, Quick also provides this thing called an activity feed. So we launched a number of items today with regards to that. But what the activity feed does, it goes across all these things that are happening in your email. So you could have a long conversation with like 10 different threads on something. Pulling that together in order to think about, "What is the point of view? What is your decision needed? What is the next step?" And surfacing the most important ones, especially when you have to make approvals and decisions. The activity feed ends up being a singular hub in Quick for most users today that use us. And what we're seeing is that change, you ask a question about change, is it fast? Is it slow? Once people get the hang of it, that AI fluency part that Russell was talking about, it becomes a second nature to them to do something in Quick. Now the part that's interesting is also that as a product, we've also made it extremely self-aware. Quick knows what it can do and how it can do it. The traditional model would've been you roll something out and then you teach people, "Hey, this is how you click this button, this is what you do." So again, don't go back to the SaaS way of doing things. Do it because AI is different. You've got to think about it differently. People have to understand that. That's the AI fluency part. But once they get it and once they start picking up the tool, adoption is extremely fast. Because you'll find that they find it a new way of working. It becomes a surface where they do a lot of their tasks, from approvals to sending a quick response to an email. All of that's possible from the same surface and that really helps accelerate those deployments.

Gemma Allen

>> Well, it certainly provides a new level of synthesis, right? Which I think we all need and enjoy in our lives in a productivity perspective. So folks, thank you so much for telling the story and thanks for joining us on theCUBE.

Russell Jukes

>> Great. Thank you very much for having us.

Jose Kunnackal

>> Thank you.

Gemma Allen

>> I'm Gemma Allen coming to you from the AWS Summit here at the Javits Center in New York City. We are bringing you all of the coverage on the ground, talking to some of the builders, breakers, and innovators generating this next wave of AI for AWS. Don't go anywhere.