Explain AI Like I’m Your Gran | Mark Werth

EPISODE 3: Explain AI Like I’m Your Gran | Featuring Mark Werth

Artificial Intelligence is everywhere… but what does it actually mean for business?

In this episode of Behind The Stack, Mark Werth breaks down AI in the simplest way possible — explaining how AI works, how businesses are using it right now, and what most people get completely wrong.

We unpack real-world AI use cases, why infrastructure matters more than ever, and how solutions like Hewlett Packard Enterprise Smart Choice servers can help businesses leverage AI without overspending or overcomplicating their tech stack.

If you’re a business owner, entrepreneur, IT manager, or simply curious about AI, this episode will help you understand the future of AI, business automation, machine learning, and scalable infrastructure.

Topics covered:

• What is AI?
• How AI works in simple terms
• AI in business today
• AI infrastructure explained
• Machine learning & automation
• HPE Smart Choice servers
• Future of AI in business
• How to implement AI without overspending

Listen now and learn how AI is changing business forever.

Transcript

SPEAKER_00 0:00

AI has become so overhyped that people stop listening. In this episode, we simplify AI so that even your grand can understand it. But if you don't understand it, you will fall behind. So, Mark, if I was your grand, how would you explain AI in the simplest way possible?

SPEAKER_01 0:30

Okay, so I normally use the restaurant analogy. The way I put it is you see a bunch of chefs in a kitchen. All of them have got their own background, their own recipes, their own way of cooking a meal. And every scenario within the AI environment is very specific to that situation. Okay. So each chef has got his own way of prepping its meal. It knows what he did wrong the last time, you know what he did correct. And all of them kind of sit in the same kitchen and make sure that they put down the food correctly. Make sure the taste is where it needs to be, it looks presentable, it look it smells great, and that's kind of what the AI environment looks like. It is very specific to each meal, it's very specific to each environment, and each individual chef kind of brings their own flavor, their own way of doing it. Like I mentioned, what they did wrong and what they did right. So then they all put that knowledge and that information and how to perfect this meal into a final presentable and customer kind of user-friendly setup.

SPEAKER_00 1:25

Hundreds. And that and the tech, explain the tech behind it. Like what does it look like in layman's terms?

SPEAKER_01 1:32

So I have to admit, a lot of people I don't think understand what AI actually means. A lot of people look at ChatGPT and think that's AI, that's the future. It's actually not. AI is a lot more vast than that. Okay. So when we look at the algorithms, the mathematical equations, you look at pattern recognization, there's so much data and information that needs to be computed, that needs to be a kind of transacted with before you can actually get a user-level output. So when we look at AI, like I said, there's a vast amount of information. And if you put garbage in, garbage coming out. So yeah, that's definitely the way it goes.

SPEAKER_00 2:07

And the actual like tech in in terms of servers, in terms of what's processing, like you say, that amount of information. What does it? I mean, shed some sort of like light into what is it that environment actually like?

SPEAKER_01 2:23

Okay, so when we look at what is required to make an AI environment, AI solution work, there's a lot of back-end moving components. Okay. First, you need data, you need information. It needs to be fed into a knowledge base where the AI can go and look for it, go and transact with it. Like I mentioned, garbage in, garbage output. As soon as you don't put the right information in or filter or see or recognize the correct patterns, a lot of the information that you're going to work with is going to put incorrect or garbage output. Okay. So once you've got that vast amount of data and information, you kind of need to compute it. You need to transact with it, you need to interact with it. And that is kind of where the HP portfolio comes in quite strong, quite well. You look at the high performance compute environments. I mean, that's basically a whole building or a whole room of infrastructure and making sure that data is being processed correctly. And like I mentioned earlier in how AI kind of works, is once you've got the data, it needs to be transacted. Each time the data is transacted, it kind of remembers the last time it interacted with it and how it interacted with it. And you need now to store that answers or that information that's been processed. So again, you look at your storage environments, you need to scale rapidly. Okay. So again, if we look at our portfolio of storage in the Electro environment, that makes it quite scalable. You can scale up, you can scale out seamlessly. So that's very important to make sure that once you've got the data and you've processed it, you can store it. And then nothing of that is going to be relevant if you don't have the software to interact with it. Okay, so we look at platforms where you kind of need to interface with it, you need to manage the AI mathematical algorithms in the back end. So software is very important. And then after you've got all this moving and everything is in the day-to-day operations, how do you secure it? How do you make sure that there's no ransomware tax? How do you make sure that your data stays relevant? So that is a very important part in making sure that your AI environment or your high performance computer environment stays relevant.

SPEAKER_00 4:13

Will we keep scaling? Will we keep having to like increase data centers and stuff like that? I mean, information we've like you're saying, we've got to capture, we got a hold, and it's got it's gonna keep growing. Yeah, 100%. So like where do we find the the space to almost have infinite amount of information living somewhere?

SPEAKER_01 4:33

So I have to admit that is why I love the HP portfolio. Again, we're giving you a lot more for a lot less. We look at the HP's compute environment where it can consolidate a lot of the legacy kit, the older generation compute environments to a lot less. So previously had bulky, clunky type of compute environment, storage environments. Now they're giving you a lot more with a lot less of rackspace requirement. So, yes, you will need to ever grow and ever expand. There will always be a demand for it. There always need to be data more stored, a lot more data transacted with. So, again, that's where HPE's, I mean, the new Gen 12s on the Proliant uh series is is quite good. That I mean you can consolidate a lot of legacy servers all the way back to just one or two gen Gen 12 nodes. So again, HPE is aware of how they need to make sure that they don't stack a whole room full of of uh kit. They kind of want to consolidate it and give you a lot more output with a lot less.

SPEAKER_00 5:31

And businesses that are that you seeing that are using AI and they're using this type of technology well, and it's it's making a difference. I mean, that's that's the thing. We want tasks to be simpler, we want work to be more seamless. Are there guys that are winning at it?

SPEAKER_01 5:46

Definitely. The the one the the one scenario I always use, and I think that's one thing that us as the humankind can be really proud of, is in the medical field. I mean, either you use medical imaging to take x-rays and scans and kind of see preemptively, is there something wrong? Is there going to be something wrong in the future? Does the certain patterns that I recognize look like it's gonna be an issue in the future? And I think that for me is from a human perspective, from going forward, that that is for me is quite impressive. I mean, if you just look at the HP portfolio, if you look at the Aruba Central, which has got built-in AI, if you look at intent-based provisioning, if you look at day-to-day task automations, a way of securing, you look at data style console, like I mentioned, the storage, automating the environment, kind of freeing up your your day-to-day guys' uh time. You don't want them running around. You kind of want if you can automate certain tasks with with the right corrective information, you can kind of free up some space and free up some some resources.

SPEAKER_00 6:37

For businesses out there, is is adapting or including this in their current scope or in their environments simple? Is it challenging?

SPEAKER_01 6:48

No, look, AI is is a funny conversation because a lot of people think they need it because it's going to make life easier, they're gonna reduce headcount, it's gonna be automating of tasks, life is just gonna get better. But honestly, you need to understand your environment first. Is there actually a set place for it? Uh, is your workloads, is your data center environment ready for it? Again, if you need to expand rapidly to kind of allow for this computing the high performance compute that needs to happen. And again, like I said, you need to make sure that your infrastructure and your workload and your day-to-day business requirements actually need it. So you need to understand the environment to know whether or not it will fit because there's a lot of back-end stuff that needs to happen. You need to feed it the information. I mean, buying the kit is the least of your worries. Making sure the correct information is fed, making sure that it transacts with your environment correctly. That that is the key portion of it. So there's a lot of pre-scoping and preemptiveness that needs to happen with this type of conversations before you just rapidly start deploying AI. I also look at when you drive to certain coastal areas, you see half-built houses or you go into security complexes, you see a guy around out of money mid-project. So that's kind of where the AI kind of sits if it's not done and scoped correctly. So yeah, be careful of rushing into the AI conversation or the high performance compute conversations and making sure that you understand and you've budgeted correctly for it.

SPEAKER_00 8:07

Solutions like you mentioned the HPE and SmartChoice. How do they handle kind of some of the workloads in different environments and their integration with different software? Is it seamless? Are they helpful for even handling? Like we're speaking about some of these AI workloads.

SPEAKER_01 8:26

100%. There's sometimes a little bit of a misconception as what Smart Choice, the Smart Choice Purchasing Program is. Smart Choice is all about taking what you know is HBE as the compute, or whether it be storage or tape backup libraries, and taking that exact same kind of build that you would require and the standards that you expect from HP and just bundling it up in SKUs. It's not third-party, it's not refurbed, it's not bought or back-end stuff. It's all the still standard HPE kind of way and standards to it, but it's just bundled up in SKUs. So yes, 100%. There is the option of adding accelerator and uh graphics cards into those specific nodes. So yes, you can actually utilize that smart choice purchasing programs for it. Like I said, if you do the conversation around project-based build configure to order, I mean that you kind of need to sit down and build everything from scratch. Simplifying the process towards it is going to make it easier. So you take a certain set of bundles that will do exactly what you need, you add some additional components within to it, and it ships rather quickly.

SPEAKER_00 9:24

So then obviously, Mark is a technical team lead at Axis. What are like what would you suggest for for people listening, for everyone out there? Like, where would you say start? Start the AI, start integrating and looking at things like HPE and and solutions like their smart choice at any level. What is your kind of advice to people jumping into it?

SPEAKER_01 9:47

I would suggest start off a conversation with resellers or the the service providers is go understand the environment. Go look at what they actually need. Do they ful need a full-blown deep machine learning system? Do they need a gentic AI? Do they what kind of solution will actually fit into the environment before you bargain with the solution? Don't help you put something down that's not future-proof, uh, that's not sustainable. So when you look at that, understand it first, then go look at what's going to be the right fit for you. Do you need to integrate multiple softwares? Do you want one kind of vendor solution? What kind of setup are you actually looking to get from what you're going to purchase?

SPEAKER_00 10:24

And some of the challenges you've seen, or whether it's environments or businesses, what are you seeing people struggle with out there and and obviously speak to how they can avoid it or things to look out for?

SPEAKER_01 10:36

So I have to start off with people that's got a big misconception about AI is. Like I mentioned earlier, it's not a it's not a chatbot, it's it's all about decision trees, it's all about uh uh making use of the algorithms to to get to the the answer or get to your outputs as quickly and cost effective as possible. So every time you kind of run a script or you run the algorithm, you kind of work and bash your environment a little bit. So every time you kind of request something, your your system gets a little bit degraded. So that's when you make sure whenever you request something, it happens up front. And if there is if you have encountered with a certain scenario, you kind of just jump back to it. It doesn't need to re-recalculate everything from scratch.

SPEAKER_00 11:13

A bit of a random question then, but obviously with the whole AI exploding, I think in many different ways, what is it gonna do to jobs, like realistically? Like I like I'm saying randomly, but for me, chat is my personal trainer, my dietitian now, like email check. So, what happens with with all these things to to people out there that are concerned at different levels? Because there's so much that it can do, so much information that we wouldn't be able to process or or be able to do. Like, yeah, I think it's just interesting, it's obviously super exciting, there's a lot of opportunity, but on the flip, what does it do to people's jobs?

SPEAKER_01 11:54

I I don't think we're close to the age and time. Like I said, if you look at the gold rush we're sitting in at the moment, I mean a couple of years back it was the cloud rush. Everybody had to get their uh their environment off-prem. They need to get it on cloud, they need to experience it, they want to spin up, spin downwards seamlessly without any issues. But if you look at the AI conversation again, as soon as people start realizing exactly you know what the the issues and the hiccups and the growing pains were with cloud, they kind of started migrating back to the environment, which causes quite caused quite a bit of quite quite a bit of issues. So back to the AI conversation, I think it's pretty much the same thing. You need to be very sure and very pedantic on how you want that AI environment to interface. I don't think people really need to get worried about losing their job or needing to close down a business because everything's gonna be automated out. There's still a human factor behind what people need to interface with. So, yes, it can automate some things, it can make life easier, it kind of recognizes the patterns and being preemptive and being as quick as possible. But from what we're seeing currently in the market and what we're seeing in the environments, I don't think that's yet to be a concern. Like I mentioned, cloud environments was the big rush at some point. So everybody has ran to the cloud environment, everybody wanted the kit off-prem. But I mean, at the back end, they they needed to come back because they realized it's not gonna fit or it's not gonna work for that environment. And that's specifically with AI as well. Um, the AI gold rush that we're seeing at the moment is people are jumping onto it, expecting to reduce headcount, expecting to automate a lot of tasks, save a lot of money. But ultimately, like I mentioned earlier as well, if you don't feed it the correct information, if you don't utilize it the way it should be, then you're gonna sit with half-built house that you you can't complete.

SPEAKER_00 13:31

Yeah, and then it doesn't improve the business, it doesn't change the thing. Very interesting. What should we look forward to with AI? What should we be excited about, about what it can do for businesses, for us as individuals?

SPEAKER_01 13:45

Like I mentioned, the HP portfolio is quite fast. And I mean, for helping your day-to-day operations team, that's kind of where you want to start it. You want to make sure that you free up some resources, you don't want them doing the minor stuff, you want them to focus on the bigger task. And I think with the likes of our AI ops, op sharp portfolios like that can actually help you automate those tasks, can advance your security, it can make sure that when certain aspects need to happen automatically or when certain security stuff can happen automatically, you kind of do that preemptively. So the software will handle it for you. There's no worrying need to jump onto a thing, fix it manually. Everything can happen, happen in an automatic sense. So I guess automation is a big thing. Cybersecurity is also a huge thing that we're seeing with all this vast amount of data. People need to get their hands on it. You kind of want to keep it protected. And I also believe in automating certain tasks within your environment. So if you have certain systems in place, you kind of want that to be as responsive and as quick as possible. Okay, so you definitely don't want to sit and wait for something to happen. The back end, reporting and stuff like that can happen automatically. So you kind of want that freedom of movement, and I believe that's kind of where AI is pushing it, making sure it frees up resources rather than like you asked the question earlier about job security and etc. So it's there to assist rather than to displace.

SPEAKER_00 14:59

And if you were to give like three words or a word or a couple on the the quick fire benefits of AI, like what what would you say? What would they be?

SPEAKER_01 15:12

Definitely the security portion of it, automation and freeing up resources.

SPEAKER_00 15:17

Super helpful. Anything we've left out or or something that you would like to leave everyone with, like in light of our conversation, in light of AI and and the way the world's going and the speeds at which we're operating. Any any final words?

SPEAKER_01 15:33

Yeah, like I said, don't be scared to interface with it. I think a lot of people don't understand it. Within the HPE portal, there is a lot of ways you can kind of interact with some of the AI tools that we've got. Um, I think it's very important to understand it before you start selling it. So for me, I would like like what said if go to the go to the HP portal, go have a look, go have a feel, go play a little bit with it, understand actually what we're positioning and why we're positioning it, and I think you'll see the great benefit of it.

SPEAKER_00 16:00

Mark, thank you so much for your time. Again, huge, huge value, super insightful. A lot I need to take time to study, understand more, and I think all of us. But I think like you say, it's exciting to to keep learning and what's gonna happen with with all this technology. So thank you.

SPEAKER_01 16:17

Thank you, Jake. We're quite quite excited to see where this AI market is gonna go.

SPEAKER_00 16:21

Thank you again.