Woodson Martin of OutSystems appears at the AI Agent Conference 2026 to discuss agentic systems, legacy modernization, governance and strategies for controlling artificial intelligence. theCUBE Research records the session in partnership with NYSE Wired. The conversation examines OutSystems' transition to an agentic systems platform, the Mentor AI coding agent, rapid modernization of legacy applications, hybrid deployments, and real-world implementations in banking, insurance and public sector environments. Martin describes the strategic role of platform-level governance, hybrid deployment models and digital sovereignty. They highlight OutSystems' approach to accelerating legacy modernization through agentic tooling and partner networks and emphasize cost control through model hot-swapping and the use of distilled or open-weight models. Analysts note implications for profit and loss predictability, vendor leverage and long-term operational resilience.
Host Gemma Allen of theCUBE Research leads the discussion. Key themes include platform governance, hybrid deployments, developer community scale, machine learning operations and low code development. The session provides practical strategies for managing AI inference costs and operationalizing agentic systems in regulated industries such as banking and insurance.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For theCUBE + NYSE Wired: AI Agent Conference
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for theCUBE + NYSE Wired: AI Agent Conference.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: AI Agent Conference
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: AI Agent Conference. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Tal Carmi, WalkMe
Woodson Martin of OutSystems appears at the AI Agent Conference 2026 to discuss agentic systems, legacy modernization, governance and strategies for controlling artificial intelligence. theCUBE Research records the session in partnership with NYSE Wired. The conversation examines OutSystems' transition to an agentic systems platform, the Mentor AI coding agent, rapid modernization of legacy applications, hybrid deployments, and real-world implementations in banking, insurance and public sector environments. Martin describes the strategic role of platform-level governance, hybrid deployment models and digital sovereignty. They highlight OutSystems' approach to accelerating legacy modernization through agentic tooling and partner networks and emphasize cost control through model hot-swapping and the use of distilled or open-weight models. Analysts note implications for profit and loss predictability, vendor leverage and long-term operational resilience.
Host Gemma Allen of theCUBE Research leads the discussion. Key themes include platform governance, hybrid deployments, developer community scale, machine learning operations and low code development. The session provides practical strategies for managing AI inference costs and operationalizing agentic systems in regulated industries such as banking and insurance.
In this interview from the AI Agent Conference in New York City, Tal Carmi, chief information officer of WalkMe, joins theCUBE + NYSE Wired's Gemma Allen to discuss how enterprises can close the gap between AI ambition and real workforce adoption. A pioneer of the digital adoption platform category and now an SAP company, WalkMe has spent a decade helping organizations extract measurable value from their technology investments — a mission Carmi says has grown more urgent with AI. He outlines a dual-layer approach pairing top-down mandates with bottom-up emplo...Read more
exploreKeep Exploring
What does WalkMe do, and how does it help companies improve user adoption of technology and realize greater ROI—especially with the rise of AI?add
How should an organization approach AI adoption—balancing top‑down strategy from management with bottom‑up use by employees—and what challenges and guardrails should be considered?add
How wide is the gap between executives' and employees' perceptions of the effectiveness of workplace digital tools and AI, and how can contextual AI (e.g., WalkMe/Gemini) help bridge that gap?add
How can a company increase employee awareness and adoption of AI tools by integrating contextual AI into their everyday workflows?add
>> I'm Gemma Allen with theCUBE and NYSC Wired, and we are here at the Agentic Studio at the AI Agent Conference in Manhattan where we are talking all things builders, breakers, buyers, and bots for this next wave of tech. Joining me now is the CIO of WalkMe Tal Carmi. Welcome, Tall.
Tal Carmi
>> Hello, thank you.
Gemma Allen
>> So for those not familiar with WalkMe, maybe let's just start with a quick, who are you, what do you do?
Tal Carmi
>> So WalkMe, we are an SAP company. We provide a digital adoption platform. Well, we have been the first in that category and been leading it for 10 years. And what we basically do is try and bridge the gap between people and technology. I think that would be the best way to describe it. How do I get people to adopt tools? Which obviously now AI is probably the most revolutionary tool we've been having forever maybe, but that's what we do. We try and get companies to get more value out of their investment because in general, many companies are not seeing the ROI they would've expected out of the tech stack. And with AI, I think that made it even worse because the adoption challenge is real. And it's even... And I think I've just been to the conference and you hear, doesn't matter who is speaking, in some way or another, everybody's speaking about adoption, the challenge.
Gemma Allen
>> So as a CIO, the value and the efficiency and the productivity, it starts at home, right? You want to make sure that within your own organization, you guys are what you sell, I'm sure.
Tal Carmi
>> Yes.
Gemma Allen
>> You're acquired by SAP, a huge company, I mean, it's monolithic really. I'm sure the integration at this time is interesting, especially in the world of AI where a lot of productivity tools, a lot of the typical tech cases or the stack from 10 years ago is somewhat right up under discussion as to what it would look like 10 years from now. Talk to me a little bit about how you approach that as a CIO. What are you buying for? What are you planning for?
Tal Carmi
>> So as a CIO, and right now WalkMe, we are operating like I'm operating as independently, right? So our tech stack is independent, right? So it's not necessarily related to the SAP tech stack internal. But in general, the way I... I have no idea how it's going to look in 10 years. I don't think anybody knows even how it's going to look in 2 years, let's be honest.
Gemma Allen
>> You're right about that.
Tal Carmi
>> Two years ago, nobody, it's like, "Oh, we have this new cool ChatGPT thing." And now we're like, "Oh, agents and what can we do?" Nobody knows. But in-house, we have been doing a lot of AI initiatives. We are very... Everybody says they're an AI-first company, but we really are. We have been pushing internally, again, leveraging a variety of tools, including WalkMe our own, of course, to try and get the maximum value we can out of AI. And it's been a journey, to be honest. I think right now, a lot of companies are trying to understand, okay, how do I get AI to actually have an impact on my business? And there is a lot of fluff out there. And it's very... If you look at LinkedIn, everybody's fully agentic 100%. Nobody's working anymore, but reality is very different. And I think there are two ways to approach it, and I think you need both of them. So you have the top down, me as management kind of saying, "Oh, we need to use AI and let's choose what we are going to do with AI." And then I have all the field. I have the actual employees that I want them to use AI for their day-to-day, which is bottom up. And I think you have to have both layers. If you only say, "Oh, use AI, use AI." Okay, some of them will, some of them will not. Some of them will use AI that you don't necessarily want them to use because you have your own certified stack that you want people to use. And if you go bottom up alone, you might get a lot of little wins, but not necessarily anything strategic or anything that really has an impact. So that's the way I see it. And again, and adoption plays a key role in that again, right? That's why I'm saying that's the biggest, I think, challenge is how do I let people, first of all, use it to allow people to play with the right guardrails, allow them to innovate, create, because that's what AI does. It really makes it much easier for whatever idea you have in your head to materialize it, which something that would have taken a month earlier can be done much faster now, can be done by individuals. On the flip side, you have a lot of people who are sort of getting lost. They're not sure what tools they have. They're not sure how to use them.
Gemma Allen
>> So when we think about that first touchpoint, you have a big team, let's say practice area, a marketing team, and you say to them, "You guys, we want you to explore this. We want you to play around with this." Typically, in previous tech waves, that would be tied to a hyperscaler, like Microsoft released Teams or by Slack or whatever, we start to embed these technologies and they're sticky. And we've seen cases whereby they're sticky for decades. When we think about this new world of kind of frontier LLMs, how sticky do you think this technology is? Are you going to have staff playing around with Codex, playing around with Claude? What sorts of discoverability is happening at the enterprise and at the business level?
Tal Carmi
>> You're right. Stickiness is a problem right now, because I don't know if you've seen this meme, you have the big players and then you have this little star saying, "Oh, you are here." And it's like, who is the best model? And you guys just stick, "Oh, right now you're here."
Gemma Allen
>> It's changing every day, right?
Tal Carmi
>> Yes, but next week you might be here. So it was like, "Oh, OpenAI or ChatGPT." And then it was like Claude. And then Gemini came over from Google and like, "Oh, Gemini is the next thing." And then it's like, "Oh no, forget about that. Now we are talking about Claude Desktop." And I'm like, "Oh wait, have you tried the new Codex?" And not even speaking about the open source model that are out there that keep on popping like Kimi. So it's a cycle and DeepSeek, of course. So it's a cycle. So if you keep on switching, it's kind of pointless. So at some point you're saying, "Okay, I'm going to stick with this for now, maybe in a year's time or half a year, I will make a change because everything is moving so fast." But the concepts are sticky. Are the tools really sticky right now? No, because you would use Claude and it would have, and it just came out and it was Claude Desktop and you had Claude Code. And then you're like, "Oh, that's amazing." Everybody's been using it. It's like, "Oh, now Claude Desktop or Cowork." So you can actually have files and you can have non-developers use it in a way that is very easy. And then they brought out skills. So people were trying to use it and now you got this ecosystem, which is very Claude and like, "Oh, this is amazing. It's mind blowing." And now you have Codex supporting exactly the same concept. You're like, "Oh, if I need to migrate, okay, Codex can work with my skills."
So if you call that a harness in a way, like Claude Desktop is a harness, Claude Code, Codex, I don't know. So tomorrow there might be somebody. And Microsoft has been pushing, saying, "Oh, why don't you use our foundry, which is cloud-based, you can run agents and we are LLM agnostic." You choose, you can switch.
Gemma Allen
>> What do you think as a CIO and as a buyer, I'm sure of technology, is the kind of predictable and constant sticky point? Is it, for example, integration? Do you think about this stack and you think, well, we already use, say for example, G-Suite, so we might as well try Gemini.
Tal Carmi
>> Yes, of course.
Gemma Allen
>> If we can make this work. What do you think is really going to be the ultimate decider?
Tal Carmi
>> I think Google is, and Microsoft in a sense depends on what shop you are, but I think they are very... Because if Gemini comes, say, "Oh, I'm already connected to your Google Drive, and to your Gmail, and to your calendar." And your native, the data is not moving anywhere that it's not right now. And sometimes from pricing point, well, it's included or it's very low price. So they are very well positioned because in the end, what we're seeing, it seems like some sort of commoditization of everything, everything sort of becomes... Somebody, Claude brings something out and then whatever, a month later or quarter later, Codex now has the same capability and now Codex has something new. So tomorrow Claude will have the same. So at some point it's sort of leveling out and you're like, okay, so now is it becoming a pricing issue? And I'm sure you heard about that everything Anthropic has been going through for the last few months with the token and the pricing and everything getting out of hand and some people actually switched saying, "Okay, I can get the same job done much faster, much cheaper, much, much more reliable with whatever, provider X." And that's a state for now, a month from now it might change again. So the stickiness, I'm not sure where it is because the models are getting better, but they're catching up to each other.
Gemma Allen
>> It's tokenomics, right? There's no two ways about it. But when you think about buying a technology and having it embedded into your stack and ensuring that there's a predictability on cost.
Tal Carmi
>> That's very hard.
Gemma Allen
>> That's a tricky space right now. We have CFOs that come into the NYSC and sometimes say to us, "I feel like I'm a CTO because I need to understand this tech at a financial level that I never really needed to do before." I'm sure from your perspective too as a CIO, you have to think about buying cycles and predictable finance in ways that you've maybe didn't have to in the past. How do you weigh it up?
Tal Carmi
>> It's hard because I think definitely for us, we are like a very SaaS-oriented company. You're used to the per seat and you know the price and as long as you know your headcount, you know the cost. And now it's not. And I think the world is moving to a consumption-based model, right? I think this is driving everything. Salesforce just came out with a headless announcement right now, it's more marketing than actual any change, but it will become at some point. And I think more companies are moving to consumption-based, which changes everything in how you budget. And I think just like you had the rise of FinOps in the world of R&D, you're going to have it now for internal CIOs because it's the same challenge. So you work around it, you work with limits and you work with some guardrails of dashboard and monitoring and all of this is a cost side. What I think is still the challenge is like, you know what? Even if I put the cost aside, let's assume AI, we want to use AI, we understand that's the future. How do we get the benefits out of it? Forget the cost. Okay, let's assume we don't have a cost issue, but how do I make sure people are actually using it? And it's hard because you have two very, I think it's a very large range. You will have people that are out there, they're doing cutting edge, everything new that comes out in Claude, they already know they tried it, they burned through tokens, but they're also doing amazing things. On the flip side, you have people that are not even yet versed in asking whatever your ChatGPT, Gemini, whatever, just saying, "Hey, read my emails and tell me what's the status of the deal with X or give me a summary of the status of this customer because I'm going into a meeting."
And it's hard because it's fine that you have these edge guys doing crazy things. Again, does it have value? Does it have value that is worth the cost? I'm not sure. Maybe yes, maybe not. But on the flip side, how do I get these people to use AI more because I want them to leverage AI. I gave you tools, but if you're not using them... And you have the in-between, right? So true story, I woke up a week ago and I got a Slack message from a VP in my company saying, "Hey, I need an API key to Slack." I'm like, "Okay." And she's not a technical person. And I'm like, "Why do you need that?" I'm like, "I don't think she had ever seen the word API key before." I'm like, "Why do you need an API key?" It's like, "Oh, I'm building an agent." I'm like, "Interesting. What for?"
And then eventually it was just, she was really trying to do something very cool. She just didn't know that Claude has a native connector to Slack and weirdly Claude didn't tell her. It just went, "Oh, I'll write some code for you that you can run and it will get you what you need, but I need an API key for it." So there is also this gap and you want to allow people to run forward and on the flip side, you want to have some guardrails, so not everybody.
Gemma Allen
>> So on one side, you have the risk of inertia, people having access to a tool and you paying a price that isn't being utilized. On the other side, you have people potentially at home uploading documents to an LLM, right? There is a super user and then there is a non-believer. When we think about the super users, what are your thoughts on the risk of that? How are CISOs or folks even in your own company from a security and governance perspective really monitoring for that? People using these LLMs, using classified information from their Outlook or whatever it might be at home to just make their lives easier. I mean, we're all guilty of it, right? It's tempting.
Tal Carmi
>> Yes, I agree. And I think the shadow AI is much bigger than shadow IT, but it's the same root cause. If you have tools that are good enough, people... Because nobody wants to go rogue, right? People do that because, like you mentioned, it's very tempting because I can do all of this, which in unsanctioned tools, which I cannot do with the tools I have. So first of all, try and get them the tools that are hopefully cutting edge. And if not, at least they are as good as... So they don't feel that, "Oh, you make me ride a horse, but there is a car that is not sanctioned." And I think that there is a very big gap also, and you're right, always between what? We have an annual report called the state of digital adoption, which we have been running for a few years. And one of the things that to me stick out really one in the report I saw, it was 80%, or I don't remember the number, 80%, 70% of executives thought that they are giving amazing tools to their employees. But if you ask the employees, the number is like, I don't remember, I think 20% or something. And then when you ask them about AI, how good is AI doing really complicated jobs? Executives like 80%, yes, it can do it. Then you ask employees and it's like 8% believing. I don't know where the truth is, but it's definitely somewhere in the middle. And part of that is sometimes employees don't even know what they have. So part of what we do in-house, how we use WalkMe, right? We try to give you contextual AI where in your flow of work. So example I gave, you can ask Gemini today, "Oh, based on my emails, give me the status of this meeting and give me talking points for my next meeting with them." It will, and it will give really good results. When I showcase that to AEs, just ask them, "Hey, open your Gemini, connect it to your Gmail and type this prompt." They saw the results and their head exploded, but a lot of times they just didn't know that you could do that. And even after I showed them, a large part of them just forgot about it because we're humans, right? So you're used to the way you work. And then we did the thing that if you're recognizing an email from a customer, we would just sort of ping you, "Hey, do you want to summarize this, get a summary of the status of the deal?" Now, it's a button, it's right there. You don't need to go to Gemini, or Claude, or anywhere and type that. You just click on it and it gives you the answer and not surprisingly, you'll see much better adoption. So it's a lot of things. So it's not just the tools. It's like you mentioned, how do I get people to actually use the tools? We built an internal... Everybody does it now, like an internal sort of rug engine. You can ask it, "What is the status of this customer?" It will pull data from everywhere and will give you a status or give me a SWAT analysis, prepare me for the next QBR, whatever you want, and it will do it. But many people either didn't know this tool existed. And even if it did, they would just go and use like, "Oh yeah, they told me it can do the status. So give me the status of account X and it will." But why aren't you using the rest of the capabilities we have? They didn't know. And then we just use, us internally, we use WalkMe, but to sort of show them like five or 10 different options, you click on it, all it does is take that information, plugs it into the tool and press enter for you, right? And all of a sudden people used it.
Gemma Allen
>> Well, I guess, it's the age-old problem, right? People process technology, but Tal Carmi, it's certainly a fascinating time to be a CIO at this mad moment where we are building things that we probably didn't even know would be a possibility 10 years ago. Thank you so much for coming on theCUBE. Great chatting to you.
Tal Carmi
>> Thank you very much.
Gemma Allen
>> I'm Gemma Allen here at the Agentic Studio at the AI Agent Conference in New York with theCUBE and NYSC wired. We're talking all things builders, bots, buyers, and breakers with the next wave of tech. Stay tuned.
Gemma Allen