Welcome, and thank you for standing by. I would like to inform all participants that this call is being recorded. Parts of this call may also be reproduced in JP Morgan Research. If you have any objections, you may disconnect at this time. I would now like to turn the call over to James Morgado. James, please go ahead.
Thanks. Thanks, Sam. Welcome everyone, and thank you for joining our AI Tech Talk with Insight Enterprises and JP Morgan. I'm James Morgado, Senior Vice President in Finance, and CFO of Insight North America. Leading the discussion today will be Matt Jackson, Insight's Global CTO, and Joseph Cardoso, Vice President, Equity Research, at JP Morgan. Before I turn the call over to Joe, let me get some of the legal stuff out of the way first. As a reminder, all forward-looking statements that are made during this conference call are subject to risks and uncertainties that could cause our actual results to differ materially. These risks are discussed in greater detail in our most recently filed periodic reports with the SEC.
All forward-looking statements are made as of the date of this call, and except as required by law, we undertake no obligations to update any forward-looking statement made on this call, whether as a result of new information, future events, or otherwise. Now, with that out of the way, I'll turn the call over to Joe. Joe?
Thanks, James. Good morning, everyone. Today we have Matt Jackson, Insight Enterprises Global CTO and SVP of Solutions, to discuss AI applications as a follow-up to our Friday's Tech Talk on AI infrastructure. Before we get started, I also have a couple of housekeeping items. First, I just want to point you to JP Morgan disclosures that Sam, our operator, just placed in the chat room or chat function on Zoom, or separately, you can find it on jpmm.com/research/disclosures. Please, when you have the opportunity, I would encourage you to view it. Second, if you have any questions you'd like to ask Matt, please feel free to use the Q&A function in Zoom today. I will be more than happy to read them on your behalf and get them answered for you. With that, let's get started.
Matt, you know, first off, thank you for taking the time today. We had a great conversation or discussion with Juan on Friday, and looking forward to having one today.
Yeah, looking forward to it, and thank you so much for having me.
Yep, of course. You know, I'm sure folks are very familiar with Insight, maybe you can just start off by giving us a quick background on yourself and your experiences, just to kind of set the stage here.
Yeah, absolutely. Happy to. Started my career, hands-on keyboard, software developer, working with enterprises to modernize applications. This is back in the 1990s and 2000s, kind of during the big, you know, tech bubble then. You know, have really evolved my career by helping to both, you know, build solutions for clients, but also building companies and organizations and teams. Started a company in 2010 called BlueMetal, which was acquired by Insight in 2015, and since that point, really been helping Insight transform from the inside. Building out our services capabilities, really launching advanced services like app mod, Data and AI, Cloud. Then, this year took over as global CTO, so setting the strategy for our portfolio and our offerings, basically what we're doing for our clients, and trying to make sure that Insight stays ahead of, you know, new and innovative technologies.
That definitely sounds like an interesting journey, and congrats on the promotion.
It is. I think I have the best job in the company, for sure.
Obviously, it's even more interesting now, right? We, especially going to this year, obviously you probably have had your head in it a little bit earlier than we have.
Mm-hmm.
AI obviously has been a hot topic this year.
Yeah
Particularly from an investor standpoint, right? It's obviously getting a lot of attention from the investment community over recent months. Maybe you can just start off by outlining what's the key difference between traditional AI and generative AI?
Mm-hmm
positive from our perspective, obviously the generative AI is the new and hot topic.
Yeah.
maybe we can just level set and just kind of parse out the details between the two.
Yeah, absolutely. You know, AI is not new. I think, you know, you go back to some of the earlier days of computing, really in the 1960s is when, you know, the foundation of AI was established, or at least the theory. It's evolved, you know, continuously ever since. You talk about traditional AI, we think about things like machine learning, you know, which is really training models based on large data sets to do, you know, fairly straightforward predictions, to deep learning. That's when you start to introduce, you know, neural networks to do more predictive analytics. You can start to get, you know, larger data sets involved.
We've been using these for years to help, you know, companies do predictions around financial markets, you know, risk around insurance, you know, customer buying patterns, supply chain issues. You know, AI has been core to what we've been doing for a number of years. You know, the difference, though, between all of those technologies and generative AI is really the scope of what is being influenced. If you look at machine learning, you know, or even deep learning, they're typically trained against a single use case. It could be a complex use case. Like I said, it might be, you know, We did a large project for a healthcare company, trying to estimate the length of stay that somebody would be in the hospital so we could get better, you know, activities around treatment plans.
Very complex, huge data sets, but ultimately it was a very narrow scope. We're asking it to solve one problem, giving it one use case with a data set and an expected outcome that we're looking for. Where generative AI really kind of broke the mold is that it can handle a wide variety of use cases. Now, it's still mostly use cases. You know, generative is an important part of the naming here, because I think people think it can do everything, that it's all of a sudden artificial general intelligence, and I'm happy to talk about the differences there. What I really try to inform clients and people that are interested in this is, it can handle a wide variety of use cases where you're generating content.
Could be texts, right, emails, you know, I'm sure we're all using it to help, you know, ChatGPT or whatever, to help us write things. Images, there's a ton of stuff out there to help create AI-based images. Where we're seeing a ton of productivity is around software development and code or even test data. And so any of these scenarios where you're looking to generate content, whatever, almost whatever form you're looking for, you know, it can handle a wide variety of these use cases. I think that's the fundamental difference, is both the breadth of what it can do, not a single, but many use cases. As well as its ability to really generate content that mimics what humans would generate. It's really the closest thing we have seen to mimicking human behavior and human creativity.
That's interesting. Obviously, I wouldn't be surprised to know the answer to this question before I ask it, but it seems like that would be just incremental, like, you just see incremental interest from customers.
Mm-hmm.
I guess, like, what are you hearing from customers today around their interest in leveraging generative AI?
Yeah.
And are you actually seeing customers using it today? If so, what for? Like, can you provide us some examples of where you're seeing customers deploy it?
Yeah. Yeah, it's interesting, and I've, the demand is incredible. I actually had. In the last month, I've had one client meeting where we didn't talk about generative AI. We, we all came out of the meeting, and we're like: Wow, that was a first in a while. It was actually an advanced networking project. It was really interesting, but, we just didn't happen to touch on generative AI. In the last month, I was keeping count of the number of clients I've spoken to, either in kind of large group settings or one-on-one, and once it got over 100 clients in couple of weeks, I kind of lost count.
We've gotten feedback and questions from a tremendous variety of clients, from the largest enterprises to, you know, small and medium-sized businesses. You know, wherever we see one of these, you know, kind of revolutionary technologies, you know, hit the market, you know, there's usually this kind of there's the early adopters and the people that wanna wait and see. People wanna do an ROI analysis and find out, you know, what the impact's gonna be. The number one question I get from virtually every client with generative AI is: How quickly can I get it? Not, not how much is it gonna cost, not what's really the ROI, what's the use case?
It's just, how quickly can I get an environment set up where my users can start to leverage, especially the chat capabilities, the generative text capabilities, in a secure and private way? I think that's the key at the end there, is that, you know, and anybody can go and sign up for whether it's Bing or whether it's, you know, the OpenAI version of ChatGPT, you know, or Bard, or you name your public, you know, generative AI capability out there. Most of the companies and most of the clients that I'm talking to, they wanna leverage that, but in a private, secure environment. They know their users are using it. They know that they can't prevent this from happening, even if they come out with the best policy in the world.
They wanna find out how quickly can they stand something up internally that they can give to their users to kinda take them away from the public sphere and get them working internally. I can talk about all the different use cases that we've seen, but, you know, I think what's most interesting is just, you know, that they are so excited about the productivity gains that they're gonna get from these capabilities. You know, and there's further use cases that we're exploring. Right out of the gate, I think it's just people have seen an immediate impact. I think it because it hit the consumer kinda sphere so quickly and just exploded. It was, what, the fastest to 100 million users of any technology ever.
And so people just wanna get that productivity gain and get that capability in-house as quickly as possible. From there, we start to explore other capabilities, but, yeah, we've already started to deploy this. We've had it internally for, I don't know how long, weeks, if not a couple months now. We've got an accelerator to get our clients using it. We've already started to deploy it with clients. We're using it, but we've also started talking about more advanced use cases, which get really interesting.
Maybe the other side of the coin there is what we're hearing is concerns around generative AIs, right?
Yeah. Yeah.
Particularly customers viewing it as maybe a disruptive or a disruption to their business.
Mm-hmm.
I guess, have you seen any customers express that concern? Like, who do you think, from your perspective, is more at risk around generative AI disrupting their business practices? I guess, who's better positioned to benefit from generative AI, at least from your perspective?
Yeah. Yeah. Well, there's, like, three questions there, I'll break it apart. I think the first one, let's just start with the immediate risks to almost anybody that uses this technology. We've done polling, with hundreds of clients and gotten some really good feedback. We're gonna launch it in a few weeks. What we found was, the number two concern, number two, that clients had was security, right? Data privacy. We thought that would be far and away number one, but it was actually number two in the polling we did. It's a very, I mean, when I say number two, it was a close second.
It's a big concern, and that's, I think, why you see so many people that wanna, you know, stand up these private instances, either on private infrastructure, as you probably talked to Juan about, or even within, you know, one of the hyperscalers, but ultimately in a private tenant, where it's their data, and they can control that. Privacy and data security is really important. The number one risk that people had was really, quality of work and the impact to human creativity. What they're afraid of is that as people start to use this technology more, they're gonna become overly dependent on it, and it will lead to a degradation in the quality of the work. I mean, you've seen some of the articles that have been written by this stuff.
It does mimic a human, but is it really as good as the best human, right? So both in terms of the output and what that might do to the quality of either the customer experience or the employee experience, or over time, you know, one of the biggest risks is just that people become overly dependent on it, and it stifles creativity and innovation. So, you know, that was surprising. I actually found that, it was good that people were thinking about that. I think that was the right focus. Yes, security and data privacy, that's like a checkbox you have to check because it's a huge risk if anything happens there.
I was impressed with kind of the thoughtfulness that people were applying to the broader topic of what does this do to humans and how we think and how we work together? That was the number one concern. I think that, to answer part one of your question, really, any organization should be, you know, focused on those areas. You know, how do you introduce this functionality without stifling innovation or creativity? How do you do it in a secure way? The second question, though, is really who is going to, you know, potentially be disrupted by this? I think that's where you look at industries where there is generation of content, right?
I know, within our organization, one of the biggest questions is, if this makes developers 50%-100% more productive, what does that do to our business model, developing solutions for clients? In our case, we came to the conclusion that there is such a backlog of app modernization and enterprise app dev that has to happen, that this actually improves our value prop. We can do things faster for clients, work through their backlogs faster. I doubt we're gonna run out of applications that need to be built or modernized, I think in other industries where you maybe have seen, you know, it could be, you know, creation of graphics, you know, marketing content, editorial content, you know, even, you know, over time, what this could do for animation and movies, right?
You could easily imagine a world where, you know, the Netflix is just automatically generating cartoons for your kids, you know, and there's no humans behind it. I think that there's the real concern that within certain sectors that rely heavily on content generation, that this could be a big disruptor. The last part of the question, right, who's gonna win? One of the debates that we had early on, is this a sustaining innovation? Is this something that's gonna benefit the existing big players in the space today, like the hyperscalers, the chip manufacturers, you know, people that are already in the AI space? Or is this a disruptive technology that's gonna create the next set of unicorns that pop up out of nowhere, that disrupt industries, long-standing industries?
The conclusion we came to is it's a little bit of both. You know, I think you've already seen that the level of investment that's necessary to develop this technology, to build these foundation models and these LLMs and come up with this technology is. It's massive. It requires lots of people spending lots of time, lots of compute, right? And actually, a lot of lawyers, too, because, you know, as you see this stuff rolled out, you know, especially we look at what's happening in the EU or even Canada, and I'm sure it'll hit the US eventually, the dialogue's there, but they haven't actually started the legislation. You know, you're gonna need if you're rolling out, you're gonna need teams of lawyers working with local governments and regulators.
I think that in that sense, the people that build the core infrastructure here, the Microsofts, the Googles, the Amazons, the NVIDIAs, the AMDs, you know, it's gonna benefit them, even the OEMs who are gonna build private infrastructure on top of this stuff. That's gonna be a sustaining innovation that concentrates a lot of the benefits to those hosters and the people that can kinda get to market quickly and sustain the investments that are necessary. I think it's gonna create an ecosystem where startups can build on top of that. You know, just like you saw with the internet, and actually, cloud's a great example. You know, where would some of these, you know, large, you know, SaaS providers be without cloud, right? If they didn't have an environment to build their product on.
That's where I think you're gonna see the disruption, is not in the major, you know, kind of technology providers that exist within, you know, private and public infrastructure, but you're gonna see it in the startups. Building, you know, AI-enabled applications that displace traditional applications, especially in vertically aligned applications. You might see a manufacturer, somebody that's been in the space building manufacturing software, retail software, healthcare software for years, and then somebody comes in with an AI-enabled version of that and displaces them. I think that's where you have to watch out.
If you're a kind of a legacy ISV, especially if you have kind of a niche that you're in, you know, you're probably ripe for displacement. Those are actually the companies that we've been talking to, and they're saying, "Hey, how quickly can we enable our applications to leverage this technology, right? That we can build better customer experiences, you know, more efficient workflows, you know, better designs if we're in R&D, so that we don't get displaced by the next, you know, ISV that builds on top of this technology.
No, it's super interesting. Very interesting to see the number one risk.
Yeah, I was. I mean, I found that, like, it was good to see, cause I feel like people are really thinking about the human impact here.
Yeah, it's a bit sci-fi movie-esque.
Oh, yeah. No, totally. Yeah, nobody was talking about when we reach, you know. Artificial well, people are talking about artificial general intelligence. This is the first step there. To go back to kind of the question about how is, you know, generative AI different than some of the other technologies, it's not artificial general intelligence, but because of the breadth of the use cases, the thing that's missing is the ability to set its own goals, to ask its own questions.
Right.
When that happens, and it will, we're really, you know, we've dipped our toe in the artificial general intelligence, and that's where we get into sci-fi stuff.
Yeah. No, super interesting.
We'll talk in a few years about that.
Looking forward to having you back then.
Yeah. Or just a bot version of me.
Maybe we can go on that. Follow off on that and just talk about, like, what's in the market today.
Yeah.
Like, what are you guys seeing? Like, who are the players? What are customers leveraging in terms of these generative AI offerings today?
Yeah.
like, what are they specifically addressing? like, maybe you can even touch on how you see them maybe evolving, not necessarily longer term, but maybe-
Yeah
just in the medium term, like what does it look like today, and what will it look like in a year from now in terms of what you're seeing from these cloud service providers or what have you?
Yeah. Yeah, it's interesting. I think that, you know, the most obvious thing that we're seeing in the market today is the productivity tools, right? Whether that's just the chat capabilities and, you know, whether it's the OpenAI public ones or if it's, you know, now that some of the hyperscalers are starting to roll out their own versions of those, whether they're public through things like their search engines or private within the enterprise. That's the first thing to hit the ground. People are already using it. You know, it's interesting, we surveyed our clients again, you know, over 70% of them had already come up with policies around the use of generative AI. Basically the other 30% were working on policies or intended to create them.
Everybody sees this being part of the workplace in the future. Whether you're, you know, Microsoft and the Copilot tools, you know, those are gonna, those are gonna roll out incrementally. Not everybody can get access to those right away because Microsoft's got to build the infrastructure for this stuff to scale. Google Workspace, another great example, where they've announced a lot of these, generative AI capabilities just woven into the, you know, the productivity suites and stacks. If you're writing an email or a Word document or a PowerPoint presentation, whatever it is, having these, generative AI capabilities, I think that's the first thing you're gonna see, right? That's almost immediate. Like I said, it's as soon as people can get access to it, they're gonna start using it.
As fast as the hyperscalers can roll this stuff out, I think in this space, you know, Microsoft will be the dominant, Google will be number two, and they address kind of different markets for productivity with Office and Workspace, but they're gonna be the dominant players in that space. Not that there won't be others. You know, there'll be CRM tools. Salesforce will come up with their stuff. Plenty of productivity gains that you're gonna see across the whole ecosystem. I think what gets interesting is the next wave, where we start to see AI-enabled applications. This is when companies start to leverage these foundation models, maybe build their own or tune their own foundation models or large language models. When they start to pull in their corporate data.
One of the things that we've done is, you know, we've got our, you know, InsightGPT set up, but we've started to pull in private data sources through things like cognitive search. When somebody goes in there, if I go in and I go to our private, you know, chat environment, I can say, "Hey, pull together some case studies that include this, that, and the other thing." It'll actually go out there and pull together those case studies. It'll look for our private data and pull it all together in a very secure, you know, private environment. You know, you're gonna start to see people layer in their own data, layer in their own models, and then ultimately, I think you're gonna see applications get launched. These could be, you know, new SaaS-based applications.
They could be industry-specific applications. They could be internal enterprise applications that leverage this stuff natively. I think that's where you're gonna see, you know, I mean, the productivity gains you're gonna see through even just Office is gonna be tremendous. I think you're gonna see some real game-changing applications get launched in the next three to five years. You know, I think about what we do within healthcare and, you know, the ability to predict, you know, medical imagery, right? To help diagnose, you know, cases that maybe you couldn't do before or manufacturing and the ability to come up with new product designs that people wouldn't have even imagined, because it's impossible to understand the whole set of data that you might need to interpret to come up with a new design.
I think as compute gets better, you know, obviously, NVIDIA, you know, has come out strong at the beginning here, others as well, around GPUs. As, as that compute gets, you know, loaded into the hyperscalers, and you can run these more advanced models and train them and embed them in applications, you're just gonna see whole new use cases, you know, pop up that we couldn't have even imagined before. I think that's kind of how it'll roll out, and like I said, I think the hyperscalers have a great play there. You know, I talked about Microsoft and Google, but all three of them have great API-level services for this stuff.
Amazon as well is coming out with almost like a marketplace, where you can buy foundation models from, you know, Stability AI or other providers, so they're basically creating this ability to kind of ingest other, you know, foundation models. Google has a whole library of foundation models that they've created that you can plug into. If you're a manufacturer, if you want to evaluate supply chain issues, they've got models that are already tuned for that type of stuff. Obviously, Microsoft, with the OpenAI investment, probably hit the market first and has the strongest offering out of the gate. They're gonna be the winners, but I think that what you'll find is everybody will benefit from the productivity stuff, but the ones that really invest in embedding this into their core applications are gonna benefit the most.
Now, you brought up an interesting point at the end of the day. It's like the delivery, righ Mm. Amazon Marketplace in terms of.
Yeah
Them trying to deliver that. You know, that kind of leads into my next question, and it's: how important is the channel during this? Mm kind of technology cycle, right?
Yeah.
I guess, one, like, how do you see Insight's role or maybe even broader, the channel in terms of dispersing this to a broader ecosystem
Mm-hmm.
You talked about having all these meetings where everyone's bringing it up to you.
Yeah.
I think what's probably even more interesting is if you can, like, look back at prior technology cycles, and maybe there's not one that's an apples to apples example. I guess the easiest or low-hanging fruit would be cloud generally, right? Maybe you can just talk to, you know, has there been an incremental focus on the channel being able to deliver this? Because it seems like that's where the, maybe the air pocket is everyone's hearing about this, but they don't know how they can leverage it in their business today. Do you see kind of incremental weight pushed on, of the Insights of the world in terms of bringing this out to customers and kind of showing them what the capabilities are and how they can leverage it?
Yeah, absolutely. Like I said, you know, we've had more conversations on this topic than any other, you know, disruptive technology that I can remember in my career. Whether you talk mobile or cloud or. You know, I came in a little after the internet, but even if you go back to that, you know, kind of the nineties, or even earlier. I think when we look at this, you know, there's a lot of. It's not just the technology implementation. Like, there's varieties of complexity, right? If you just want to stand up your own private, you know, chat instance, that's one thing. You know, eventually, Microsoft through, you know, O365, is just gonna enable the ability to upgrade to get Copilot capabilities. You know, our teams are already leveraging GitHub Copilot to help write code today, right?
Seeing the productivity improvements. I think there's some things that are low-hanging fruit, but what you have to remember is that this is not just about co-deploying technology, right? It's about understanding the impact to your organization, and I think that's where we at Insight, we're obviously in the channel. The investments that we've made over the years in real services capabilities, both technology services as well as things like, you know, product management, organizational change management, this is disruptive to companies. When you introduce this technology, it raises questions about, you know, what level of employment do you need? Do you need X number of people in these different offices anymore, or is it just a productivity improvement?
You know, is this going to power, you know, exponential growth within your organization because your existing team is, you know, so much more productive? Those are not things you want to find out accidentally, right? Those are things that you want to be thoughtful about, that you want to map out a roadmap to introduce this technology. You want to understand, you know, how it's going to change workflows within the organization, how it's going to change the way that people use technology, right? Your enterprise systems, your third-party systems that you've stood up. I think there's obviously a channel play there in helping..
You know, if the value there is really helping clients understand, you know, how they can get access to this technology, who they should go with, how they should deploy it, we're a critical component to that thought process and that kind of decisioning process. I also think that there's this, like, when we talked about risks, there's a human impact discussion that needs to occur. There's a security and data privacy discussion that needs to occur. I think the channel is critical, and in particular, where we can deliver value-add services, advice, and guidance to our clients as they adopt this technology, that's the real differentiator. That's what they really need. They need access. They need somebody to help them deploy it.
They need somebody to help them figure out which provider to go with based on their, you know, kind of unique use case. More than anything, they need somebody that's going to hold their hand through the process, that's done it before, right? We've already done it ourselves. We've done it for clients already, we can talk to them about the pitfalls, what to watch out for, you know, how to proceed, what the best practices are, even just designing a policy around this type of stuff. We're really, I think, at the early stages of that, but like I said, it's picked up faster than, you know, anything I've seen before. You mentioned cloud, that's probably the best analogy. I mean, mobile might be.
You know, everybody that had a website had to create a mobile version of it as soon as, you know, the iPhone was launched. So you see these kind of S-curves of innovation, where you have this disruptive, innovative technology, and then everybody rushes to adopt it, and then it levels out to the next one. This is definitely one of those S-curves. So we have to be there to help our clients through every step of that journey, right? Because it's not just turning it on like a light switch. It really is thinking through, you know, what are the use cases? Where does the right technology apply to. You know, like I talked about earlier, you know, it's not every use case.
It works best for use cases where, you know, people are actually generating content, generating things, so identifying the technology that can support that, identifying then how the organization is going to change to adopt and get the best use out of this technology, and then ultimately deploying the technology. There's a lot of phases to get this right. I think that's where, like I said, channel, and especially channel with services, comes in to play.
No interesting. I guess on that note, you know, where's Insight today relative to making generative AI solutions available for its customers? I guess the buts of my question there is, you know, the way that I kind of think the go-to-market rolls out, and you can correct me if I'm wrong, is probably you partner with customers on projects that you guys are kind of working on in tandem with, and then as soon as you kind of end up creating some type of solution, you probably look to basically bundle it and almost make it like an off-the-shelf kind of offering, right?
Yeah.
I guess if I'm correct in characterizing that, you're like, where are we in that journey? Are you seeing more of the, you know, R&D/experimental phase still, or are you guys actually having those bundled offers today because you worked on some early-inning projects, and so therefore, you do have some off-the-shelf offerings today? Can you just kind of elaborate on that?
Yeah, well, I think you hit on kind of the. I lead our portfolio team and how we launch things and bring them to market. Ultimately, you know, I think of it in a number of different categories, but I'll focus on a few. Packaged offerings, where we have some IP, where we've done it before, we've got a runbook, we've built efficiencies around delivery, we're leveraging our offshore teams, we're leveraging that IP both to deploy as well as to run. Maybe we have managed services around it.
We develop these packaged offerings, and they can either be, you know, targeted towards the scale commercial segment, where we can deploy them hundreds of thousands of times, or they can be starting points for larger projects, where we customize on top of them, but ultimately, that's kind of one category. The other category is where you start out really from an envisioning, right? There, there may be reusability, there may be assets, there may be IP, but it might be a very unique use case for an enterprise, right? They're trying to do something for the first time that nobody's done before, so there's not going to be a, 100 or 1,000 examples of it.
Those are more of the, you know, custom engagements, where we do deep evaluation of the use case and kind of map out that journey for them. We're doing both of those right now with generative AI. We've already built the accelerator. We've got it in some of the marketplaces today, like on Azure. We're able to deploy a generative AI environment for a client in four hours. You have to get permission from Microsoft. A lot of the stuff is in pilot, so there's, like, sometimes a two-week wait to say yes, but once they say yes, we can get this stuff stood up in a private tenant for a client, so they can use these chat, you know, generative AI chat capabilities in half a day.
We're doing that today for clients. We're obviously leveraging it internally, and we've been working on that for months. That's where we saw this coming. We saw it, you know, when it hit the consumer space. I mean, it was obvious that it was going to take off in the enterprise, and so we rushed to get a really solid framework in place that we could then deploy to multiple clients, and, you know, we were client number zero. I feel really good about what we've got there, and that kind of answers the question that clients have of: How quickly can we get that secure environment set up?
That is also the basis for these enterprise conversations, where they say, "That's great, you know, but we want to pull in this data source and this data source, and we want to do prompt engineering to, you know, change the logic to fit with our, you know, custom, you know, decision-making framework, or whatever they have. It could be an insurance company that wants to model risk. It could be a banking company that wants to, you know, estimate, you know, market performance, you name it. That same framework that we built to be highly repeatable and deployable at scale is actually a starting point for having a broader implementation or conversation around other custom use cases. It could even be the middle part of a custom user interface on top of generative AI.
Those are the conversations that I think will take longer, and there's a development life cycle, and it could take months or quarters or years to get some of these products to market. Like I said earlier, I think the first one, the packaged offering, is that quick hit productivity gain, and we're well positioned to help clients there. The other one I think is gonna lead to some really sustained, you know, changes within just how people work, how people interact with technology, how you interact with the companies that you know, work with, buy from. I'm excited about some of those journeys that have just begun in that space.
When you think about the strategy around evolving Insight's portfolio around.
Yeah
all these different offerings, packaging, et cetera.
Mm-hmm.
You know, like, how are you tackling that? Like, how do you see it evolving as you go into the future? Then maybe, like, a second part of this, like, who are the key partners that are dominating your share of mind?
Yeah
as you kind of are looking to evolve your offerings?
Yeah. Well, this is where I think about, you know, what is the value proposition that we bring to our clients? How is that unique in our space, right? I think Insight is. If you followed Insight, you know we've been going through this transformation from, you know, that channel partner reseller to really a global solutions integrator. For us, I think what's important is, you know, we need to be able to address our clients entire need in this space. Obviously, knowledge of the hyperscale environment and the cloud providers and what they're offering, whether it's productivity, whether it's, you know, some of this, you know, kind of the APIs that enable custom models, whether it's the foundation models themselves. Having deep knowledge and partnerships with those hyperscalers is critical.
There's not a lot of organizations that have both that, as well as the relationships with the, with the OEMs and the chip manufacturers that are going to power both the cloud infrastructure, I mean, NVIDIA is selling like crazy into the hyperscalers and AMD and others, and Intel, but also the OEMs that are going to build private infrastructure, whether you're running this in a data center. I mean, think about, Juan probably talked about all the data centers that need to be modernized to, you know, load up more GPUs than CPUs or whatever, you know, infrastructure is needed to run these models.
You know, not to toot our own horn too much, but, like, I don't think there's another company that has the depth and breadth in both the hyperscale cloud environment as it relates to generative AI, as well as more of the private infrastructure and OEM environment, as Insight, right? There's nobody that has that heritage of both worlds, and they're really necessary. If you think about the ecosystem that exists today, you know, all of the big players, you know, fit into our category of top partners. For us, you know, we're having conversations with each and every one of them about their product roadmaps. We're designing reference architectures.
When a client comes to us and they work with a particular OEM and a particular cloud provider, and they want a hybrid environment, so they can run these generative AI workloads in the cloud, in the edge, at the core, we can help them do that across that whole technology landscape. Whether it's, like I said, product roadmaps, reference architectures, or actually working to co-invest in developing IP, this accelerator that we've got in market right now is built on Azure OpenAI. We can spin that up very quickly. We collaborated closely with Microsoft, both to develop that as well as to bring it to market, getting in front of clients, evangelizing its capabilities.
I think that type of motion is going to be key to landing this offering, but I really do believe. You know, I can't think of another company that's got the breadth of capabilities, the channel, the partnerships, the deep services expertise around cloud, app dev, data, infrastructure, to really solve these complex problems for clients.
Sounds super interesting. You might have touched on my last question, but it's a two-parter, so maybe the second part will be a little bit unique. Before I get to my last question, just want to remind folks, we'll definitely hit the questions that are in the Q&A section after I finish mine, but obviously, if you guys have any other questions, please feel free to drop it in the Q&A box, and I'll read it on your behalf and have Matt answer it for you. Yeah, just, you know, you kind of touched on this already throughout the conversation, right? You've mentioned Microsoft a couple times. I don't think it's a surprise to anybody that Insight has a very close partnership with Microsoft.
You know, as you look forward in the context of generative AI, you know, how are you seeing that relationship develop? You know, do you think it's gonna be a tighter and closer relationship between the both of you as you kind of lean on each other to get these generative AI solutions out there to a broader customer? You know, the second part of my question is: Is there anyone in the ecosystem that could almost be like another Microsoft, right?
Mm-hmm.
Obviously AI, Microsoft was originally developed, or that relationship kind of was developed on this concept of getting cloud out there. Even if you look at your competitive landscape, there's some folks that are, that are considered the large VARs of today, right? They're not even leaning into cloud as much as some folks might even think of, right?
Mm-hmm.
From that context, like, how should we think about the ecosystem? Is there another play into AI? Like, obviously, the low-hanging fruit here is like a NVIDIA relationship.
Yeah.
Is there another, like, tight-knit relationship that folks should kind of think about in terms of as this, the relationship develops, that maybe they're not really considering because it's not typical for the channel to have that partner in that ecosystem?
Mm-hmm.
You know, just trying to get any thoughts around that.
Yeah, absolutely. There's a couple questions there. I'll start with the Microsoft relationship and where we see, you know, this strengthening that relationship. Obviously, you know, with them being really first to market around, you know, their partnership with OpenAI and launching the OpenAI capabilities within Azure, as well as what they've done with Copilot, you know, that's already accessible within GitHub for developers, and it will be coming out, you know, in the various, you know, productivity suites, Office 365 over time. They've probably got the most buzz in the market, you know, that they've kinda came out of the gate as the dominant player with the first mature offering.
As one of Microsoft's largest partners, and Microsoft obviously being one of our largest partners, you know, we've been working with them for a long time in this space so that we could both capture it. The demand exceeds their ability to capture it right now, or to fulfill it, I should say. You know, we're working with them in the commercial space, really their scale, to build these offerings, you know, like our IP, that accelerates the adoption of OpenAI and Azure. You know, they are.
I won't say they're struggling, but they have such kinda inbound demand from clients, that they need partners like us to help satisfy that, to meet with the clients, to talk about what the product does, to talk about the right way to deploy it, to add in that IP and the services to make sure that it's a successful deployment. They can't do that on their own. So when they see this type of demand, especially in that, you know, kinda SMC space for Microsoft, they need partners like us to help fulfill it, and we're perfectly established. We've been doing it for years, and we can scale up to support that. Especially with some of the investments that we've made, you know, building out our teams offshore, a lot of this delivery can happen that way.
We can build these package services and deploy them quickly to get clients up and running, you know, on the Microsoft Azure, OpenAI, and Copilot capabilities. I think right out of the gate, I've talked to, you know, global sales teams at Microsoft. We've done events in the field. We've brought clients in. You know, we're doing a roadshow. I think this is just gonna deepen our partnership there. I think in the enterprise space, we're also a really large Microsoft partner. That's where you're gonna see a lot of co-investment in big client initiatives over the next coming years. You know, Microsoft's fiscal starts in about two weeks, and we're already talking to clients about, you know, what those really large initiatives are gonna be.
In their, you know, top 500 accounts, you know, that's where you're gonna see major transformation projects, digital transformation projects, to modernize entire, you know, suites of tools within the enterprise. That's where I think also our partnership comes in strong, is that we've got really deep technical expertise, not just to help them scale, I mean, that's really important, but when they're picking their biggest account, that's gonna build the most complicated enterprise system that leverages this technology, we're the partner they come to. I think that's really important in strengthening the relationship as well. They've gone through a lot of changes. I think that, you know, they always go: You know, are they more internally focused? Are they more partner-focused?
Obviously, they know they've got a great product right now, but they've been great to get, you know, in conversations about what the partnership looks like, how do we go to market together? Cause Microsoft knows that partner ecosystem, Insight at the top of that list, is really gonna help them scale and capture this opportunity. I'm excited about that. I think that is only gonna strengthen and grow our Microsoft relationship. You asked about, you know, kinda who else in the space to talk about NVIDIA. There's a lot of complementary partners. I think that a lot of the chip manufacturers, I mean, you know, a lot of these workloads, you know, drive consumption of NVIDIA product or AMD or even, you know, Intel and others.
You know, a lot of the companies that might have been a step behind, this is such a big market, they're gonna catch up. We talked to partners that, you know, might not be as far ahead as NVIDIA or Microsoft, but, you know, we look at their product roadmaps, and they're gonna get there, right? This is a big enough market that I think there's space, for all of the established players, assuming that they're thinking about it and they're innovating and they're investing, to play a big part in this. We talked to the chip manufacturers, but the OEMs, you know, Lenovo, Dell, et cetera, that we've got great relationships with. You know, there is gonna be.
I think this, and Juan probably talked about this, the amount of investment that's gonna happen in modernizing data centers, in building, you know, GenAI appliances that can run at the edge. You know, there's a lot of use cases that we see where people are gonna wanna run these models on private infrastructure. It might be connected to the cloud, it might be a hybrid, you know, it could be Azure Stack, it could be, you name it. The, you know, whether it's highly regulated industries, whether it's, you know, critical national infrastructure, defense, there's gonna be so many use cases that light up that whole ecosystem, where I think it especially, and it's not an either/or, it's not cloud or edge or data center.
It's an "and" conversation, and the more flexibility you can create in moving workloads, you might train it in the cloud, run it at the edge of the data center. You might train it in the data center, run it in the cloud. You've got to look at cost, you've got to look at security, you've got to look at just the practicality of what you own and where you want to invest. I think that's where, that's where Insight, you know, thrives, right? That's where clients come to us for that type of advice. I think that ecosystem is going to be strong. I don't think there's a lot of people that can address it, but I know Insight can.
I, you know, Just because Microsoft was first, and obviously, you know, they're our largest partner, you know, almost all of our enterprises are multi-cloud. They tend to use different clouds for different workloads. You know, they're not necessarily taking one workload and moving it across multiple clouds. What I've seen that's really interesting is, you know, there's a little competition there. I talked about, you know, Microsoft with their productivity suite and Office 365 and Copilot and Google Workspace and what they're gonna launch. I actually think each of the cloud providers has come up with something a little unique, which could create opportunities in this multi-cloud or many cloud-type environment. Microsoft, obviously, OpenAI. Everybody knows, you know, what that can do, and there are other investments that they've made in other forms.
They kind of put it under the umbrella of Copilot, you know, it's the, it's the code, it's the email, it's the presentations, it's all that type of stuff. I mentioned Google has actually built out a library of foundation models themselves. Google is known for their expertise in data, right? A lot of our clients use them, you know, for cognitive services, data workloads. I think they've got a really compelling and somewhat different offering from Microsoft, where they've started to invest in building custom models that really are only accessed via APIs. This is where you want to create a custom app or a custom process.
I think they've got a unique offering there. Like I mentioned, with Amazon, you know, they're obviously, you know, if you know Amazon, they're a great, you know, large marketplace, and that's one of their strengths. They've started to differentiate by pulling in, you know, other, you know, like I mentioned Stability AI, Anthropic. There's other providers that have built these models, other kind of ISVs and third parties, and they're kind of pulling them in under this umbrella, so you can get access to those models within the AWS environment. I think all of the hyperscalers are gonna thrive. Some might be a little ahead of others, and obviously, our relationship with Microsoft is extremely strong. I think you've got to keep an eye on all of them.
You've got to look at the chip manufacturers, then, you know, ultimately, like I said, I think there's an OEM play. I didn't even really talk about software, right? I think, like, we're probably gonna run out of time. I want to leave time for the other questions, but I've talked a lot about the hyperscalers and the infrastructure providers, but within the software space, I mean, you talk about CRM tools. You know, is there gonna be a CRM tool that doesn't have, you know, generative AI? Is there gonna be an ERP system that doesn't have generative AI? Is there gonna be, you know, any type of third-party system, whether it's healthcare, retail, that doesn't have these capabilities embedded in it? Within five years, I don't think there's gonna be any major platform that doesn't have this.
We don't have time to go through all of those, but obviously, Insight in the channel, you know, I mean, Adobe, you name it, right? Everybody that we work with has this on their roadmap. I think it's an exciting time to be in the channel. It's an exciting time to be a technologist, and it's an exciting time to be leveraging this technology.
That's super interesting. you know, as you said, got a couple questions here for you.
Yeah.
We'll start off at the top. You know, the first one reads: How will the data stack evolve to GenAI trend? Will Lakehouse players like Snowflake and Databricks proliferate? Will NoSQL players like MongoDB become more popular among developers?
Interesting. I won't compare and contrast, you know, Databricks versus Snowflake versus whatever, but I will tell you that when we look at these solutions, the number one pull-through, so not necessarily the productivity one. Well, actually, maybe even the productivity ones when we're doing these, you know, chat capabilities and pulling in enterprise data. The biggest pull-through offering that we have behind generative AI is modern data estate. What we're doing is we're talking to clients, and they're saying, "Hey, how quickly can we stand up one of these, you know, chat capabilities, either in our corporate intranet or through a tool like Teams?" Our answer is, "Hey, you know, hours, not days, not weeks, not months. We can stand this stuff up in hours." They ask, "Okay, but then how do I pull in my corporate data?
How do I pull in either stuff that's in my enterprise, you know, data mart or, you know, data warehouse?" Like you talked about, like the Lakehouse's and, you know, Like, all of these capabilities to modernize the data estate, are going to see increased interest and adoption. There are some questions about, you know, do you, do you really need to cleanse all your data, or can generative AI effectively allow you to skip some of the steps? You know, master data management, right? Like, you really need to create perfect alignment between all your different, you know, data sets and normalize all your identifiers, or can generative AI figure that stuff out for you? That's a really good question, right?
Like, if you have two different systems that have Matt Jackson. If I go to a hospital, and I've got Matt Jackson in two or three different systems, you know, that's a big problem for healthcare today, and there's a lot of work that has to go into aligning those. Can generative AI maybe alleviate the need to do that? What does that do to some of the software stacks that you just talked about? I think that's a legitimate question, but regardless, the investments that need to be made in modernizing data and making data available to these models probably outweighs, you know, any potential impact or decline in the complexity of implementation, right? It becomes easier to do this in some ways, but it becomes more necessary as well.
Yeah, interesting. Another question from a client goes: Where does your AI value prop. Okay, let me rephrase this. Essentially, they're basically asking you, how does the competitive landscape look like relative for you guys versus your peers? Mm. Yeah, basically, that's how it reads.
Yeah, absolutely. Well, I think three things. The breadth of our capabilities and partnerships is pretty unique. I talked about that quite a bit, so I don't think there's other companies that have both the cloud partnerships and expertise, the OEM, the chip expert. Nobody else that has that breadth. I think there, we definitely have a really unique value prop. If you've got a complicated environment, you know, regulatory, technology, you name it, even from the human workflow factor, you know, I think we're well prepared to have those conversations. I also think that, you know, we are very, we're all in on this, so we have invested in ramping up our teams.
I won't say all of our developers are using generative AI because, you know, we need to work with clients to make sure they're comfortable with that, and there's risks involved there. Where our clients are comfortable with leveraging that, where we have, you know, agreements in place, our teams are already leveraging that, and we're seeing productivity gains. For basic software development, we're seeing productivity gains, and Microsoft published, I think 56% is what they saw through a study using GitHub Copilot. We can validate that. I we don't have that level of precision, but we're definitely anecdotally seeing those types of gains. For things where you're modernizing a code base, you're basically saying, "Hey, we're gonna upgrade this code base from X to Y," we're seeing 10-20x improvements, right? Orders of magnitude improvements in productivity.
I think that's another area where we've got a value. When we're engaging clients and modernizing an application, migrating a data source or you know, data estate, modernizing the data estate, we can now do that faster than we could before. Sooner or later, people will catch up because it's gonna become table stakes. You're not gonna win these projects if you don't have that level of productivity amongst your team. But I think, you know, there's a window here where the faster that you can introduce these capabilities and get your clients comfortable with using them, your value prop, you can do more for less, right? Or you can get more done faster than you could before. I think that's an area where we're differentiating.
Like I talked about, kind of the, putting the emphasis on the human impact as well, and just understanding, you know, what does this do to your product strategy? What does it do to your organization's, you know, change management strategy? I think we've got some unique offerings there as well. Kind of attacking it from all sides, but I think, you know, Insight definitely has a lead in preparing to help our clients transform around this technology.
No, for sure. The next question here is: What's the most common first enterprise application that Insight is working with clients on it? Is it contact centers? Is it cybersecurity? Is it document processing? Is it unlabeled data management? Is it process automation?
Yeah.
What Copilots are they mostly implementing?
Yeah. I won't say most, but I'll say. I'll give you an example of a few. We look at all the different use cases, the basic one is actually just standing up a private chat environment so people can, you know, say, "Hey, write me an email that says X, Y, and Z," because they might not, you know, depending on how quickly Copilot rolls out, they might not get to Copilot for months or a year or more. Just standing up an environment where they can allow their employees to go in and enter in data and get responses back in a ChatGPT-like environment, that's the basic. You'd be amazed at how, you know, how much productivity gain our clients can get from a very small investment there.
That's number one. When you look at individual use cases, you know, we look at, is it more verticalized or horizontal? I'll say right now, we're looking at more horizontal use cases. The vertical will come, and there's some there, because the demand is so great, addressing some of the vertical use cases, you know, document generation or give you a good example, document processing. You know, reviewing contracts for, you know, policy violations or, you know, terms and conditions that are missing. You know, that's a low-hanging fruit. You know, you've got a library of documents, you've got standard policies, terms and conditions. Validating that can accelerate, you know, contract reviews, contract, you know, development, can definitely ease the burden on some of the legal teams.
HR, huge use cases there for employee onboarding, training, again, you know, creating agreements, creating job recs, recruiting. These are the areas, kind of these horizontal, looking at the businesses. If you've seen, you know, some of the organizations that have come out and said, "Hey, where are we gonna, you know, maybe drive productivity improvements, you know, or maybe, you know, look to, you know, hire less over time?" It's a lot of these shared functions, you know, legal, HR, recruiting, et cetera, where we're seeing some pretty big productivity improvements and just, you know, rapid adoption. I think the verticalized ones, specific to healthcare, manufacturing, retail, we're having those conversations, but it's going to take a little bit longer to build the solutions, train the models, et cetera.
Yep. Interesting.
Yeah.
The next one I have for you is: Do you have examples of startups that are bringing AI-enabled software to the market and disrupting established players?
It's interesting. Established players is tough. I won't name a whole laundry list of the startups and names in here, but if you look at, like, what we're seeing around, you know, automated assistants and bots, you know, all of us went through this journey of building chatbots five years ago, right? The chatbots of five years ago are nothing compared to the chatbots today. You look at, you know, where you can actually have, like a, you know, what appears to be a person that can actually converse back and forth and answer questions in a very natural way.
You're starting to see more, you know, investments in actually, like, hand gestures, facial expressions. That's definitely disrupting, like, the customer service applications that we were building in the past, the chatbots that we were building in the past. That's actually one of those horizontal use cases, is replacing chatbots, especially for, you know, websites, product recommendations, customer support. We're definitely seeing disruption there, where, you know, new players are coming in and displacing folks that maybe have legacy or traditional, you know, customer support platforms. You know, helpdesk is an area, right, where we're coming in, same thing, where you can displace a lot of the companies that, you know, maybe were providing helpdesk software or even personnel. You know, you can see a lot of disruption there.
You know, this isn't where we play as much in the in the, you know, kind of graphic design, media creation, but I'm seeing a tremendous amount of disruption there, right? Whether that's, you know, graphic design. We're seeing productivity improvements because we have people that design mobile apps and design websites, and they're leveraging these technologies to improve the speed and quality at which they can develop new designs. But for folks that are dedicated in that space, we're seeing a lot of displacement. I won't even get into, you know, kind of the, you know, marketing and, you know, you know, media space. Not necessarily our area, but obviously seeing a lot of disruption there.
No, for sure. Next question here says: Do you think GenAI will disrupt software business models in terms of profitability, having to invest in infrastructure/data models, negative on gross margins, having to recruit and pay less, lower on OpEx, consumption-based pricing versus subscription, et cetera? I guess, how do you see the go-to-market and kind of the monetization of that changing with this generative AI use cases around software?
Yeah. Well, it's early, I think there's some, still some things to be figured out. You know, like, one of the big questions, if we can develop code a lot faster, in the near term, that we can probably drive higher margins, right? We can deliver more for our clients for less cost, that's great for the business. Does that eventually result in pricing pressure, right? Where clients realize that these productivity gains, they want some share of that. Absolutely. That's the natural kind of adoption curve and, you know, maturity curve. I think the key is, you know, there's opportunity to be had now in adopting these technologies and getting the near-term benefit. Eventually, I think from an economic standpoint, it's gonna get priced in, right?
The value that you place on certain jobs where maybe AI can do a pretty good job replacing those roles is gonna be less, you know? I won't name some, but some have come out, and some companies have come out with announcements that they're gonna either lay off or not hire X number of 1,000s of people in certain roles. I think there's gonna be, you know, economic pressure or wage pressure on those roles. I think from an organizational standpoint, I believe this is gonna drive significant top-line growth. I think the demand is so great that you're gonna see. I mean, we're already seeing it in the market. I read an article the other day about, like, you know, why are we, you know, are we in a bear market or not, bull market?
They're like, "We should be probably in a recession, except AI is keeping us afloat," right? All of the investment that's being made in AI today, and the appreciation of companies that play in this sector, right? A lot of the ones I've talked about today, but Insight as well. I think there's an enormous opportunity to grow for profitable growth over the next few years, because, you know, there probably is wage pressure, but there's also, more than anything, just enormous productivity gains. Almost everybody across the organization, especially in tech, you know, where we're writing code, where we're building applications, you know, we're seeing productivity gains, we're seeing top line growth as a result of the investments that people are making in this space. I think there's a lot of near-term benefit.
I can't James is gonna slap me on the wrist if I give anything specific to Insight, but I think just generally, those are the trends that you're gonna see, and you're already starting to see them.
Yeah, for sure. Don't worry, I'll make sure James stays in his place.
This is broad commentary, not Insight. I mean, like, just look at the market cap of some of the places, the companies that have come out with AI technology, and, you know, the productivity gains that are, that are obvious.
I'm still listening, by the way.
Was that James? Thank you.
We were just testing. Let me just check time. Maybe there's enough time for this last one here. Are you seeing any material traction for non-LLM type AI, i.e., more quantitative focused in CPG, specifically, pricing more broadly?
Well, I don't know about the pricing question, but the demand, yes. I think, you know, there's always gonna be kind of, you know, associated demand for this technology. We'll talk to clients, and they'll say the buzz around generative AI strikes up a conversation, and we're like, "Okay, but that use case is actually better served by, you know, deep learning, machine learning, more traditional, as we talked at the very beginning, more traditional AI mechanisms. You know, or maybe the real problem that they have is that they have a, you know, distributed data estate that needs to be modernized. I think we're seeing definitely kind of pull-through demand.
app mod, data estate, like I mentioned, data estate, like cleaning up the data, but then app mod, whether that's leveraging generative AI to accelerate the pace of app mod or whether it's building new applications that expose generative AI experiences, there's pull-through for all these different capabilities. That's why I think this is like mobile, like cloud, like internet, you know, such, you know, a pivot point or an inflection point for this industry, is that this isn't just gonna power investment in generative AI, it's really gonna power investment across the IT stack.
Got it. We did perfect timing here. Let me end the call here. I know there's a couple questions left in the Q&A, so I'll get through those and circle back, and hopefully we can get Matt some, get you guys some answers, but we'll get back to you over email. With that, you know, wanted to thank you, Matt, for taking the time today. James, Ryan from Insight as well, thank you for taking the time today and setting this all up. Appreciate the time of all of you participating in the call as well. With that, we can close it up.
Great. Thank you so much. Appreciate it.
Great. That concludes today's webinar. Thank you all for your participation. You may now disconnect your lines.