Exclusive Coverage: theCUBE at IBM Think | Mohamad Ali | IBM Consulting

theCUBE.net

Clips
News
More from Exclusive Coverage: theCUBE at IBM Think

Big Blue’s AI revolution: IBM Consulting Advantage as a workbench for gen AI projects

Artificial intelligence is revolutionizing every sector, and consulting is no exception.IBM Corp.’s IBM Consulting Advantage is at the forefront of this transformation, offering tailored AI solutions that make advanced technology accessible and effective. By focusing on security, bias checking, personal identifiable information protection, governance and cost efficiency, the AI services platform is paving the way for the next generation of AI projects.IBM’s Mohamad Ali talks to theCUBE’s Dave Vellante about the importance of IBM Consulting Advantage

play_circle_outline Enhancing Gen AI Projects with IBM Consulting Advantage: Providing Security, PII Protection, Bias Checking, and More

play_circle_outline Software-led consulting and scaling the business

play_circle_outline Combination of human labor and digital labor in the future of services

play_circle_outline Goal of increasing win rate with clients using IBM Consulting Advantage

Info

Mohamad Ali | IBM Consulting

search

>> Welcome back to Boston, everybody. We're here at the BECC, where theCUBE got started in May of 2010. What a wild ride it's been. We're going to talk about IBM's Consulting Advantage with Mohamad Ali, who's the senior vice president of IBM Consulting, and he's also been named one of the top 50 influential technology leaders in Boston, actually, number 35. Congratulations.

>> Thank you.

>> I'm trying to crack the list. We'll get there, but pleasure having you on theCUBE first time.

>> Yes, first time. Thank you.

>> Great to see you. Yeah. Okay. So, the Buzz IBM Consulting Advantage. What is that?

>> Yes. So, at IBM Consulting, we have about 160,000 consulting engineers, and they're doing hundreds upon hundreds of gen AI projects. And one of the things that we realized is they need to be doing it in a consistent, secure manner. So, IBM Consulting Advantage is first and foremost a piece of software, a layer that everyone uses within consulting to do gen AI projects. It's a workbench for our 160,000 engineers. Below the Consulting Advantage layer is a whole bunch of LLMs, different LLMs, because our consultants need to be able to build solutions with IBM Granite, Llama 3 from Meta, Microsoft GPT, Amazon Bedrock, on and on and on. So, we have bolted a whole variety of LLMs underneath, almost any LLM can be bolted underneath. And then on top, our consultants are building assistance and copilots and so forth to do specific tasks. If they want to migrate something to the cloud, there is a assistant that helped them do that. And then, in the middle, Consulting Advantage provides a minimum of six things. Security. So, we're living a day now where they's something called prompt injection, where when you call an LLM, you're not sure if a malware has injected something in the prompt set and it's going to do something, so you have to secure it. You also have to protect against bad use of personal information, PII. Bias, it does bias checking. It checks for intellectual property like GPL code and other types of intellectual property issues. It helps you with cost arbitrage to pick the right LLM to do the right job. It does governance. So, now we have this layer that allows 160,000 consulting engineers to be able to build gen AI solutions for our clients in a secure and governed way.

>> So, it sounds like software-led consulting-

>> It's software-led consulting, that's right.

>> For years, obviously the big GSIs, IBM as well, you've got software expertise, you use software to make folks more productive, but gen AI has really changed the vision of a consulting and services business. How do you think about scaling such a business?

>> Well, this is actually why we built IBM Consulting Advantage because eight months ago when I looked at this, what I realized is that we had several hundred gen AI projects. And every one of these projects was being done a different way, and it's very difficult to scale that. So, by building IBM Consulting Advantage, we now have the ability to do these things in a consistent way. So, we're able to go to engineers that are doing particular tasks, like say a unit test. Let's say this group of engineers do unit tests. This other group of engineers write Python code. This other group of engineers write ABAP code. For each one of them, we could then allocate a set of very specific gen AI assistants to help them with their jobs, and they're doing it in a secure way. This way we could scale it out to 160,000 engineers. Otherwise, there's really no way to scale this out in an efficient manner.

>> When you think of the economics of the services model and versus, say, the marginal economics of software, obviously there's a huge difference. But by softwareizing the services business, a couple of things can happen. On the one hand, you say, okay, the economics are going to be much better, and especially if you're selling value of outcomes. The flip side of that is it enables more competition and more innovation. And so, how do you see the future of the business model of services?

>> Yeah. Well, I think that the business model of services is, as you described, is going to become a combination of human labor and digital labor. And in terms of how you compete, I think it's all about speed. Whichever consulting company gets there faster is going to win, and that's what we're trying to do. And we are lucky at IBM in that we have a whole technology company at our side. So, when we decide to go build IBM Consulting Advantage, we could do it quickly, we could do it with a lot of software engineers, we could do a lot of technology, and we have our colleagues ready and willing to help us do that. So, that's why in eight months, we've been able to not only launch Consulting Advantage, but it's now being used by 15,000 of our 160,000 engineers.

>> So, we've been talking about domain-specific models. Obviously, IBM Consulting has a big advantage. IBM's got a very unique advantage in it's got both the consulting capabilities, deep industry expertise that's competitive with any GSI out there. But you've got this technology portfolio product.

>> That's right. That's right.

>> And so I'm really interested in this domain specificity, and in particular, Mohamad, I'm interested in the physical and the digital worlds coming together. So, the combination of software, that industry knowledge that you have to create new capabilities. We don't talk so much about digital transformation anymore because it's all AI. But essentially, digital representations of your business, some people say digital twins, where people, places and things are occurring in real time, and that's different than the batch world that we live in. You think about supply chain, "Oh, there's a problem. I'll send some kind of message. We'll fix that tomorrow. We'll split the order." Real time is really what we see as industry transformation potential and massive economic value being unlocked. How do you think about?

>> Yeah, so I'm going to give you two examples that we're working on. One is less industry and the other one is more industry. And so, there's this particular client, they happen to be a healthcare services entity, and they wanted us to build an app for their customers to be able to sign up for appointments and that sort of thing. It's a very, very large entity. And so, we decided to use two teams. One team using IBM Consulting Advantage and all the gen AI tools associated with that, and one just doing it the regular way. And I'm not promising this for every project that we do, but the team that was using IBM Consulting Advantage and the gen AI assistants and the additional tools there, were able to do it 52% in less time and labor than the other team. So, just think about that in terms of cost savings when you do these projects, any of these projects. And that's across industry, we could apply that to healthcare or other industry. That's a generic sort of software development life cycle kind of project. Now, when it comes to industry specific, last week, was it, we made that big announcement with Palo Alto Networks, and one of the things that you'll read in that press release is a joint collaboration to jointly develop industry-specific threat models. So, there are trillions of signals that Palo and IBM now have access to. And what we're going to do is we're going to look at those signals and we're going to build, we're going to use things like InstructLab, which you've heard a lot about here, to build industry-specific models that says that, "In oil and gas, this is what the threat vectors look like in oil and gas." In financial services, the threat vectors are a little bit different. And so, then you take those industry-specific threat models, you layer it on top of a generic threat model detection capability, and then you have something that's even better. So, that's the kind of stuff that we're using IBM Consulting Advantage for and InstructLab to build models that are very specific.

>> So, what's different about that, because make sure I got this right-

>> I know there's a lot there.

>> But what struck me is you have this industry-specific threat models, you've got a signature. Does either this collaboration or even AI, maybe this is both, how does that change what was possible previously? Was it expands the accuracy and the specificity? Or is there automation or speed, because it's all about speed against the hackers, that that brings?

>> Yeah, I think a couple of things. It does improve the accuracy. You will be able to identify threats quicker that you couldn't before. And in some ways, accuracy and speed are really essential to cybersecurity because if you're five minutes late and the bad guy gets in, it's too late.

>> Or if it's a false positive, it makes your head explode.

>> Or if it's a false positive, then you have a whole ton of people triaging false positives and you just can't afford that. There's a study that says there are 3.4 million cybersecurity experts missing, so there's clearly not enough people to triage all these signals. So, I would say yes to your question.

>> And it was, I think in December that you announced this collaboration with Palo, December of last year.

>> So, that was the first collaboration, and then last week we announced this even deeper collaboration.

>> Right, right. And so, we are going to have you and Palo on theCUBE-

>> Excellent....

>> to talk about that collaboration. We'll make it happen once this conference season dies down, we'd love to do that. What about IBM Garage? How does that fit into all this?

>> So, IBM Garage is sort of the front end to this. So, you take that healthcare application that we built, there was a Garage session before that to determine what it is that we should build, not what we can build. So, why is building this application in this particular way with these features a good return on investment? It's almost like doing good product management ahead of actually building a product. And so, the Garage methodology is focused on how do we co-create with you, and then how do we co-produce with you, but ensure that it is generating good ROI? What we're doing with Consulting Advantage and takes the output from the Garage methodology and say, "Okay, here are the requirements and now we need to create these personas for this application, these user stories, these blocks of code, this set of unit tests, this set of integration tests," and so forth. Consulting Advantage is really the delivery mechanism that sits on top of IBM Garage.

>> So, we talk about systems of record and social media extended that to new ways of collaboration. People talk about systems of agency, bringing historical analytic systems of record, which has sort of been the stovepipe, and bringing transactions to that, so that the AI can actually take action. When you talk to clients, where do you think we're at in that vision coming true?

>> Yeah. So, InstructLab is an extraordinarily important tool now or invention that allows us to do that. So, today, the LLMs that we use are sort of general purpose LLMs, right? They're trained on large quantities of internet data or in case of Granite, in more curated data, but they're not necessarily specific to the data you have in your company. If you have a whole bunch of data in your company and you want to train a model on it, as of a few months ago, it was massively expensive. You have to buy a bunch of GPUs or get access to a bunch of GPUs, which are, as you know, are very expensive, at $40,000 a card. And then you have to run it and use a lot of power all the time, et cetera-

>> They're probably going up....

>> and it's going up. And so, with InstructLab, there are two inventions here. One is that you could take your company data and you could train a small model using a MacBook. Invention number one. So, you don't need all this expensive hardware. But you're training a small model. And then, invention number two is you could stitch that small model onto a large model, like a Granite, that's a general purpose model. And when you stitch them together, you could treat them as one model. And so this is invention number two. So, once you could do that every night, you could build a small model with your company data set, put it on top of this big model and have something that's actually usable, but specific to you.

>> So, follow up on that, because today, applications are generally written... And I want to make sure I understand what you said. So, we're seeing applications that use one model at a time. You said stitched together. Do you see it going beyond... Now, stitched together may mean linearly using those models, but there's talk in the application development world of actually having the combinatorial effects of those models working together, that's coming.

>> Yeah. So, there are two kinds of stitched together here, right? One is when you're stitching together a big model with a small model. And then, the other is, okay, now I've created this new model and it does a particular task, but it turns out that I need a super large model to do a different task. Let's say that different task is ingesting 90 minutes of video and summarizing it. There are very few models that can do that. But then, you have another task that you're an insurance company and you want it to write letters to your clients using your data. That's probably a much smaller model that's very specific. And we could do this in InstructLab thing of stitching a larger model with a smaller model and get to your specific thing. And one of them, one of them might cost 1 cent a token and the other one 300 cents a token. And so, what you just said is a different kind of stitching together, where if I have a copilot assistant that does this, another one that does that, I want to stitch together those assistants in a process workflow. And what we used in IBM Consulting Advantage is Watson to orchestrate the sequence of those assistants and copilots in a business workflow.

>> To Optimize-

>> Does that help?

>> Yeah, very much so.

>> There's a lot here, right?

>> In that example, you optimize on the outcome. Maybe good enough is way less expensive than perfect. And so, you can help .

>> Yeah. And actually... No, no, I'm not even saying that. I'm saying that high quality and a big model, you can also get high quality in a small model because for certain tasks, you don't need something that big. So, if you go into the supermarket, you don't need a Ferrari. You'll still get to the supermarket if you use a Volkswagen, and you'll get there just the same amount of time and quite successfully, and you'll probably bring home maybe even more groceries.

>> Deploy the Ferrari when needed.

>> You deploy the Ferrari when needed.

>> Last question. A year from now, think 2025, what do you want to be able to say that you can't say today?

>> I want to be able to say that IBM Consulting Advantage has enabled us to dramatically increase our win rate with our clients because we want to be bringing something to our clients that they can't really get anywhere else. And the faster we move to bring them a combination of human labor and digital labor, the more we're going to win. And I would hope that our business would grow. And I'll give you a quick example. So, if you think about washing machines in the home, you would've assumed that the workload, because productivity in the home, the workload in the home would've gone down. It actually didn't. From 1900 to 1960, the number of hours an American woman worked in the home actually went up four hours. So, the GDP just expanded and there was more work in the home, and there were more work in the factory. And so, I think if we move fast here, there'll be a ton more work for us, but Consulting Advantage has got to advance as quickly as we can and allow us to do that.

>> Very interesting stat. Hybrid cloud, open source, AI and consulting. Unique advantage for IBM. Mohamad Ali, thanks so much for coming on theCUBE. Really a pleasure having you.

>> Thank you. Great to be here.

>> All right, keep it right there. We're going wall-to-wall, IBM Think 2024. My name is Dave Vellante. You're watching theCUBE.

Mohamad Ali | IBM Consulting

search

>> Welcome back to Boston, everybody. We're here at the BECC, where theCUBE got started in May of 2010. What a wild ride it's been. We're going to talk about IBM's Consulting Advantage with Mohamad Ali, who's the senior vice president of IBM Consulting, and he's also been named one of the top 50 influential technology leaders in Boston, actually, number 35. Congratulations.

>> Thank you.

>> I'm trying to crack the list. We'll get there, but pleasure having you on theCUBE first time.

>> Yes, first time. Thank you.

>> Great to see you. Yeah. Okay. So, the Buzz IBM Consulting Advantage. What is that?

>> Yes. So, at IBM Consulting, we have about 160,000 consulting engineers, and they're doing hundreds upon hundreds of gen AI projects. And one of the things that we realized is they need to be doing it in a consistent, secure manner. So, IBM Consulting Advantage is first and foremost a piece of software, a layer that everyone uses within consulting to do gen AI projects. It's a workbench for our 160,000 engineers. Below the Consulting Advantage layer is a whole bunch of LLMs, different LLMs, because our consultants need to be able to build solutions with IBM Granite, Llama 3 from Meta, Microsoft GPT, Amazon Bedrock, on and on and on. So, we have bolted a whole variety of LLMs underneath, almost any LLM can be bolted underneath. And then on top, our consultants are building assistance and copilots and so forth to do specific tasks. If they want to migrate something to the cloud, there is a assistant that helped them do that. And then, in the middle, Consulting Advantage provides a minimum of six things. Security. So, we're living a day now where they's something called prompt injection, where when you call an LLM, you're not sure if a malware has injected something in the prompt set and it's going to do something, so you have to secure it. You also have to protect against bad use of personal information, PII. Bias, it does bias checking. It checks for intellectual property like GPL code and other types of intellectual property issues. It helps you with cost arbitrage to pick the right LLM to do the right job. It does governance. So, now we have this layer that allows 160,000 consulting engineers to be able to build gen AI solutions for our clients in a secure and governed way.

>> So, it sounds like software-led consulting-

>> It's software-led consulting, that's right.

>> For years, obviously the big GSIs, IBM as well, you've got software expertise, you use software to make folks more productive, but gen AI has really changed the vision of a consulting and services business. How do you think about scaling such a business?

>> Well, this is actually why we built IBM Consulting Advantage because eight months ago when I looked at this, what I realized is that we had several hundred gen AI projects. And every one of these projects was being done a different way, and it's very difficult to scale that. So, by building IBM Consulting Advantage, we now have the ability to do these things in a consistent way. So, we're able to go to engineers that are doing particular tasks, like say a unit test. Let's say this group of engineers do unit tests. This other group of engineers write Python code. This other group of engineers write ABAP code. For each one of them, we could then allocate a set of very specific gen AI assistants to help them with their jobs, and they're doing it in a secure way. This way we could scale it out to 160,000 engineers. Otherwise, there's really no way to scale this out in an efficient manner.

>> When you think of the economics of the services model and versus, say, the marginal economics of software, obviously there's a huge difference. But by softwareizing the services business, a couple of things can happen. On the one hand, you say, okay, the economics are going to be much better, and especially if you're selling value of outcomes. The flip side of that is it enables more competition and more innovation. And so, how do you see the future of the business model of services?

>> Yeah. Well, I think that the business model of services is, as you described, is going to become a combination of human labor and digital labor. And in terms of how you compete, I think it's all about speed. Whichever consulting company gets there faster is going to win, and that's what we're trying to do. And we are lucky at IBM in that we have a whole technology company at our side. So, when we decide to go build IBM Consulting Advantage, we could do it quickly, we could do it with a lot of software engineers, we could do a lot of technology, and we have our colleagues ready and willing to help us do that. So, that's why in eight months, we've been able to not only launch Consulting Advantage, but it's now being used by 15,000 of our 160,000 engineers.

>> So, we've been talking about domain-specific models. Obviously, IBM Consulting has a big advantage. IBM's got a very unique advantage in it's got both the consulting capabilities, deep industry expertise that's competitive with any GSI out there. But you've got this technology portfolio product.

>> That's right. That's right.

>> And so I'm really interested in this domain specificity, and in particular, Mohamad, I'm interested in the physical and the digital worlds coming together. So, the combination of software, that industry knowledge that you have to create new capabilities. We don't talk so much about digital transformation anymore because it's all AI. But essentially, digital representations of your business, some people say digital twins, where people, places and things are occurring in real time, and that's different than the batch world that we live in. You think about supply chain, "Oh, there's a problem. I'll send some kind of message. We'll fix that tomorrow. We'll split the order." Real time is really what we see as industry transformation potential and massive economic value being unlocked. How do you think about?

>> Yeah, so I'm going to give you two examples that we're working on. One is less industry and the other one is more industry. And so, there's this particular client, they happen to be a healthcare services entity, and they wanted us to build an app for their customers to be able to sign up for appointments and that sort of thing. It's a very, very large entity. And so, we decided to use two teams. One team using IBM Consulting Advantage and all the gen AI tools associated with that, and one just doing it the regular way. And I'm not promising this for every project that we do, but the team that was using IBM Consulting Advantage and the gen AI assistants and the additional tools there, were able to do it 52% in less time and labor than the other team. So, just think about that in terms of cost savings when you do these projects, any of these projects. And that's across industry, we could apply that to healthcare or other industry. That's a generic sort of software development life cycle kind of project. Now, when it comes to industry specific, last week, was it, we made that big announcement with Palo Alto Networks, and one of the things that you'll read in that press release is a joint collaboration to jointly develop industry-specific threat models. So, there are trillions of signals that Palo and IBM now have access to. And what we're going to do is we're going to look at those signals and we're going to build, we're going to use things like InstructLab, which you've heard a lot about here, to build industry-specific models that says that, "In oil and gas, this is what the threat vectors look like in oil and gas." In financial services, the threat vectors are a little bit different. And so, then you take those industry-specific threat models, you layer it on top of a generic threat model detection capability, and then you have something that's even better. So, that's the kind of stuff that we're using IBM Consulting Advantage for and InstructLab to build models that are very specific.

>> So, what's different about that, because make sure I got this right-

>> I know there's a lot there.

>> But what struck me is you have this industry-specific threat models, you've got a signature. Does either this collaboration or even AI, maybe this is both, how does that change what was possible previously? Was it expands the accuracy and the specificity? Or is there automation or speed, because it's all about speed against the hackers, that that brings?

>> Yeah, I think a couple of things. It does improve the accuracy. You will be able to identify threats quicker that you couldn't before. And in some ways, accuracy and speed are really essential to cybersecurity because if you're five minutes late and the bad guy gets in, it's too late.

>> Or if it's a false positive, it makes your head explode.

>> Or if it's a false positive, then you have a whole ton of people triaging false positives and you just can't afford that. There's a study that says there are 3.4 million cybersecurity experts missing, so there's clearly not enough people to triage all these signals. So, I would say yes to your question.

>> And it was, I think in December that you announced this collaboration with Palo, December of last year.

>> So, that was the first collaboration, and then last week we announced this even deeper collaboration.

>> Right, right. And so, we are going to have you and Palo on theCUBE-

>> Excellent....

>> to talk about that collaboration. We'll make it happen once this conference season dies down, we'd love to do that. What about IBM Garage? How does that fit into all this?

>> So, IBM Garage is sort of the front end to this. So, you take that healthcare application that we built, there was a Garage session before that to determine what it is that we should build, not what we can build. So, why is building this application in this particular way with these features a good return on investment? It's almost like doing good product management ahead of actually building a product. And so, the Garage methodology is focused on how do we co-create with you, and then how do we co-produce with you, but ensure that it is generating good ROI? What we're doing with Consulting Advantage and takes the output from the Garage methodology and say, "Okay, here are the requirements and now we need to create these personas for this application, these user stories, these blocks of code, this set of unit tests, this set of integration tests," and so forth. Consulting Advantage is really the delivery mechanism that sits on top of IBM Garage.

>> So, we talk about systems of record and social media extended that to new ways of collaboration. People talk about systems of agency, bringing historical analytic systems of record, which has sort of been the stovepipe, and bringing transactions to that, so that the AI can actually take action. When you talk to clients, where do you think we're at in that vision coming true?

>> Yeah. So, InstructLab is an extraordinarily important tool now or invention that allows us to do that. So, today, the LLMs that we use are sort of general purpose LLMs, right? They're trained on large quantities of internet data or in case of Granite, in more curated data, but they're not necessarily specific to the data you have in your company. If you have a whole bunch of data in your company and you want to train a model on it, as of a few months ago, it was massively expensive. You have to buy a bunch of GPUs or get access to a bunch of GPUs, which are, as you know, are very expensive, at $40,000 a card. And then you have to run it and use a lot of power all the time, et cetera-

>> They're probably going up....

>> and it's going up. And so, with InstructLab, there are two inventions here. One is that you could take your company data and you could train a small model using a MacBook. Invention number one. So, you don't need all this expensive hardware. But you're training a small model. And then, invention number two is you could stitch that small model onto a large model, like a Granite, that's a general purpose model. And when you stitch them together, you could treat them as one model. And so this is invention number two. So, once you could do that every night, you could build a small model with your company data set, put it on top of this big model and have something that's actually usable, but specific to you.

>> So, follow up on that, because today, applications are generally written... And I want to make sure I understand what you said. So, we're seeing applications that use one model at a time. You said stitched together. Do you see it going beyond... Now, stitched together may mean linearly using those models, but there's talk in the application development world of actually having the combinatorial effects of those models working together, that's coming.

>> Yeah. So, there are two kinds of stitched together here, right? One is when you're stitching together a big model with a small model. And then, the other is, okay, now I've created this new model and it does a particular task, but it turns out that I need a super large model to do a different task. Let's say that different task is ingesting 90 minutes of video and summarizing it. There are very few models that can do that. But then, you have another task that you're an insurance company and you want it to write letters to your clients using your data. That's probably a much smaller model that's very specific. And we could do this in InstructLab thing of stitching a larger model with a smaller model and get to your specific thing. And one of them, one of them might cost 1 cent a token and the other one 300 cents a token. And so, what you just said is a different kind of stitching together, where if I have a copilot assistant that does this, another one that does that, I want to stitch together those assistants in a process workflow. And what we used in IBM Consulting Advantage is Watson to orchestrate the sequence of those assistants and copilots in a business workflow.

>> To Optimize-

>> Does that help?

>> Yeah, very much so.

>> There's a lot here, right?

>> In that example, you optimize on the outcome. Maybe good enough is way less expensive than perfect. And so, you can help .

>> Yeah. And actually... No, no, I'm not even saying that. I'm saying that high quality and a big model, you can also get high quality in a small model because for certain tasks, you don't need something that big. So, if you go into the supermarket, you don't need a Ferrari. You'll still get to the supermarket if you use a Volkswagen, and you'll get there just the same amount of time and quite successfully, and you'll probably bring home maybe even more groceries.

>> Deploy the Ferrari when needed.

>> You deploy the Ferrari when needed.

>> Last question. A year from now, think 2025, what do you want to be able to say that you can't say today?

>> I want to be able to say that IBM Consulting Advantage has enabled us to dramatically increase our win rate with our clients because we want to be bringing something to our clients that they can't really get anywhere else. And the faster we move to bring them a combination of human labor and digital labor, the more we're going to win. And I would hope that our business would grow. And I'll give you a quick example. So, if you think about washing machines in the home, you would've assumed that the workload, because productivity in the home, the workload in the home would've gone down. It actually didn't. From 1900 to 1960, the number of hours an American woman worked in the home actually went up four hours. So, the GDP just expanded and there was more work in the home, and there were more work in the factory. And so, I think if we move fast here, there'll be a ton more work for us, but Consulting Advantage has got to advance as quickly as we can and allow us to do that.

>> Very interesting stat. Hybrid cloud, open source, AI and consulting. Unique advantage for IBM. Mohamad Ali, thanks so much for coming on theCUBE. Really a pleasure having you.

>> Thank you. Great to be here.

>> All right, keep it right there. We're going wall-to-wall, IBM Think 2024. My name is Dave Vellante. You're watching theCUBE.