Tom Reuner — HFS Research[00:05]
Automation is in the eye of the beholder. Put another way, it’s so important to be transparent about which lens we are looking at automation through. Is it more the business side? Are we discussing RPA? Or are we coming from the other side, the more IT-centric side, where we have different requirements, different discussions? Or what’s the next frontier with all of that? And to discuss all of that, I’m delighted to have two outstanding automation pros with me, and a really warm welcome to both Bill Lobig and Matt Lyteson. A very warm welcome, from IBM. Hope you’re well.
Bill Lobig — Vice President of Product, IBM Automation[00:36]
Thank you, Tom. Appreciate you having us.
Tom Reuner — HFS Research[00:40]
And hopefully, for the audience, you might be delighted that we’re not just discussing the RPA afterlife. Literally, that’s far too narrow a point of view. What we’re trying to do, if I boil it down, is legacy and cloud native — at least the way I look at it, and I’m curious to hear Bill’s and Matt’s view on that. It’s not a binary view of one concept supplanting another, or the narrow lens we should apply. Most clients’ requirements these days are much more complex, and I’m delighted to have such an outstanding audience with me to walk me through that. To cut through the market noise — some of that’s your job, some of it I’m just the host. So, having said all that, Bill, can we start with you? Rank and serial number — what exactly do you do at IBM?
Bill Lobig — Vice President of Product, IBM Automation[01:14]
Thank you, Tom. Bill Lobig, I’m the Vice President of Product for IBM Automation. So I define the strategy for the products, build them, support them, and we’re doing a lot of great things. I look forward to talking about it today.
Tom Reuner — HFS Research[01:40]
Great having you. Matt, same question to you.
Matt Lyteson — Lead, CIO Hybrid Cloud Platforms, IBM[01:40]
Yeah, I’m Matt Lyteson. I lead our CIO hybrid cloud platforms group. So part of my job is making hybrid cloud real, bringing value, and — for Bill’s context — using all the products that his team is building, kicking the tires on those and seeing what’s real and what’s not.
Tom Reuner — HFS Research[02:01]
Perfect. And clearly, when I said we have an outstanding panel with automation pros, just listening to you, that’s exactly what we have. So I’m delighted to have you. Without further ado, let’s jump into the topic. The discussion for this specific webinar or videocast, if you like to call it that, started at HFS a bit more than a year ago. We published a piece of research, and one question we asked as part of it was: what are the most impactful investments in an automation platform you will make over the next 12 months? For some, perhaps not surprisingly, IBM came out quite strong. And if you think about it, you guys must be a little bit like kids in a candy store given all the capabilities at your disposal. On one hand, you have the investments that are RPA-centric or adjacent. You acquired, for instance, Turbonomic to progress with the wondrous world of AIOps and observability. But you also have Red Hat, both for runbooks and the more IT-centric view of automation. And most importantly, not least for today’s discussion, you have hybrid cloud. So we’re back to the whole notion of what’s the next frontier for automation. First off, top of my mind is legacy, and integration — much deeper, not just the API integration. So we need almost much more guidance on what that really means. I’m looking forward to all your insights. So against all that background, Bill, can I start with you? It doesn’t matter how I look at or expect the evolution to happen — what’s your view from IBM? How do you look at the evolution of automation, and what is the differentiation you’re trying to convey to your clients and bring to those discussions?
Bill Lobig — Vice President of Product, IBM Automation[03:48]
Yeah, absolutely, Tom, appreciate it. To your point, RPA does have a finite applicability, and one of the things we’re focused on — and I’m going to emphasize our IT automation portfolio and the acquisitions you described around Instana and Turbonomic, combining that with our organic investments. Look, IBM is a hybrid cloud and AI company, and as we see in the industry, many clients have this strict distribution and fluidity of workloads across on-premise — to your point, legacy — to new workloads in cloud, and optimizing workload placement across those. The complexity of IT operations has exploded by an order of magnitude. It isn’t all just behind your data center anymore. IT teams need help keeping up with this, and I think it’s fair to say a lot of the tools that have been built over the years for IT operations assume that something will break and that a person needs to go fix it. Log analytics tools, observability tools — that’s important, that’s real, and that will continue. But we’re really focusing on making every computable problem automated, allowing computers to do as much as possible. And there are a lot of great examples of that.
Tom Reuner — HFS Research[05:10]
No, it’s a great point. And the way I was thinking, back to my point, automation is in the eye of the beholder: on the business side, the reference point often is an invoice, a structured piece of information; in IT it’s much more incidents. So we have to be mindful of the different requirements. But in the end, the way we think about it, and the evolving requirement of customers, is bringing those things together. Because if you don’t understand all the telemetry data underpinning your process or your delivery, some people struggle to make sense of it. But perhaps, Bill, can you help clients or the audience understand what are the outcomes you’re trying to deliver with all this innovation and this differentiation you just discussed?
Bill Lobig — Vice President of Product, IBM Automation[05:53]
So at the end of the day, the outcomes are less time spent on toil, and helping clients reduce their so-called error budget burn rate — spending more time on innovation and productivity than chasing fires and dealing with those Slack channels that have 1,000 people trying to find the needle in the haystack. We’re doing that by giving clients tools to observe, optimize, and prevent issues in their environment. Just really briefly about each of those: application performance management, sometimes thought of as observability — you can’t manage what you can’t see, so that’s where it starts, and I recommend clients get a handle on their overall estate as a beginning point. Then, optimizing that application supply chain: every piece of technology in your infrastructure needs to work harmoniously, in concert, in the right amount, at the right frequency, in the right direction, to deliver those application outcomes. You want to make sure you’re spending what you need, no more, no less — and this is particularly important with the rise of sustainable IT, green IT, and energy consumption. And then ultimately, proactive prevention of incidents, by finding, detecting, and avoiding them before they even occur — to my point about making every computable problem autonomic.
Tom Reuner — HFS Research[07:12]
That’s great insight, and suffices to say it’s a far cry from the RPA-centric discussion we often get. Everything gets masked by those discussions, even if you’re not discussing IT. It’s a struggle at times, quite frankly, but I’m delighted we have more and more enlightened discussion today. And perhaps not surprisingly, Matt, can I bring you back to the point on legacy and cloud native? Building on what Bill was saying, what are some of the insights you glean from the discussions you have with your clients? Can we go a bit deeper again on the outcomes you’re delivering, so what’s changing? But also — and this is fundamentally closer to my, or our, heart at HFS — how do we cut through all this market noise around that?
Matt Lyteson — Lead, CIO Hybrid Cloud Platforms, IBM[07:48]
Yeah, it’s kind of interesting, because as Bill was describing everything he’s hoping clients will achieve, he’s describing a day in my life. We’ve got incidents, we’ve got applications that are complex — we’ll call them our heritage applications — some of the most critical things that run our business, that maybe were built a couple of decades ago, in some cases 30, 40 years ago. And at the same time, we’ve got all these hyper-modernized, componentized applications, and I’ve got to know everything that’s going on. I want to give the right things the right resources. One of the things we found is that maybe 10, 15 years ago, application teams were really good at doing performance engineering, planning, understanding what their peaks and valleys were going to be in their workloads. And then the cloud came on. When we started to use that at enterprise scale, people got less good at that. Then we started to use some basic automation — we’ll write scripts, we’ll detect a little bit of what’s going on in that application. But I’ve got to do that across my entire estate. That’s where having more sophisticated tools where I can see what’s going on comes in — this is where things like Instana and other tools come into play, and we’re using some of that. And then I’ve got a tool that can actually help me manage that dynamically, so I can ease the burden of capacity management — things like Turbonomic — and that becomes really interesting for us. So yes, I want to see what’s going on, I want to be able to prevent any issues before they occur, and when issues do occur — because, as we’ve described, this landscape is more complex than ever. Some of the marketing stuff, because I’ve got Bill and everyone else under the sun trying to sell me stuff to help with my problems, it’s not making things simpler. It is sort of masking the complexity underneath, because the reality is, if you take two systems that are now integrated in 10 different ways because we’ve got microservice, component-based architectures, I need a way to manage that complexity and do it effectively, taking humans out of that. Having intelligent automation and the tools and sophistication to be able to do that is really what we’re after.
Tom Reuner — HFS Research[10:09]
Now, a great insight, and it’s almost a segue to the question I had in mind. On one hand, if I take the lens much more of the context you’re taking us into — so not so much the RPA business operations discussion, but the broader discussion on cloud — where organizations embark on the technology transformation. That’s rather than the cost angle, all too often, but that’s probably not for today’s discussion. If they go the next step and say we have to take cloud much deeper, it’s about business transformation. But we’ve seen all too often it’s a struggle to capture value from the investment. Part of that — and I’m building on what you said, from the outside in, being an analyst — I think we need to be much more honest. We have to articulate what the value of cloud, or being cloud native, is, but we also have to be clear that as some of my clients are embarking on that, and to be honest, operationalizing it, that’s a new complexity, at least in the short to medium term. If that’s the case, coming back to your point about autonomics, AIOps, and observability — is it just about having the insights, the telemetry data, to look at that complexity? Coming back to the initial discussion as we were discussing automation, are you somewhat providing — what we haven’t seen broadly, especially for the more IT-centered requirements — it’s tough enough to get all the telemetry data out of specific domains, make it end to end, make it an enterprise view. But then the next big challenge is to almost automate it, not just having all the dashboards saying, here’s a problem. You’re trying to anticipate. Is that where you’re heading, or is that the next evolution, probably still out?
Matt Lyteson — Lead, CIO Hybrid Cloud Platforms, IBM[11:46]
That’s where I want to get, because the reality is things are so complex, and dashboards are really great for forensic analysis. I think we’re in this state now, with the type of complexity and where some of the tooling is trying to catch up, where the reality is, yes, we’re going to need that. But we’ve, for example, in our use of Turbonomic, now taken, I think, like 45,000 actions in our OpenShift environment a month. My humans could not keep up with that. We started with basic t-shirt sizing of what an application team needs. We defaulted everyone to medium; they could request bigger, and if we saw they didn’t need as much, we gave them a little chart that would show you don’t need as much, and that got the right behavior. But then we were able to reclaim terabytes of memory, because we’re able to have the tools do that dynamically, at that rate and pace. I mean, think about it — that’s like 1,000 actions per day, more than that, that it’s performing on the environment. Humans can’t do that. And what we found was really interesting: yes, this starts out as everyone’s interested in the cost conversation. But you need to remember that’s really one-dimensional, and the sophisticated audience realizes it’s cost per performance. I could optimize costs and keep cutting costs — and typically infrastructure teams are asked to do that — but then you realize that eventually you’ll tank performance. So how do I get the right performance? What we found using these tools this way is that we start to allocate more resources for some of the workloads, and again do it in a way that humans are not able to keep up with. We start expanding that across our environment. Then, as Bill said, I can cut out some of the toil and turmoil that the teams are dealing with on a regular basis, and start to predict these things. Look, we’ve had automation for years — what’s the number one incident? Have a team go and write automation for that. We’re still going to have some of that, but as the tools get more intelligent and sophisticated, we start to allocate the right resources, we start to prevent things from happening, and where we’re headed is really what we call intelligent workload placement. I’m going to figure out the things that are important for my organization, and within the constraints, both financial and operational, I want to be able to move those workloads where they need to be. The only way I can do that is to have this foundation of intelligence and this foundation of operations where I know what I’ve got, I know what’s going on, I know whether it’s healthy or not. And I’ve got the tools that can help me ensure I’m maintaining that health — giving them the right resources or auto-resolving things. That puts me in a position where, yes, I can move them to the most energy-efficient platform, I can move them to the most cost-effective platform. As I’m streamlining my application portfolio, rather than having stranded hardware on the floor of my private cloud, I can move workloads that were in my public cloud over to private cloud and reduce some of my variable expense, making sure I’m using all my resources. This is really where we’re headed. And now, with the sophistication of these tools, we’re in a position to actually be able to do that, rather than just, let’s write some more runbooks, let’s write some more runbooks.
Tom Reuner — HFS Research[15:19]
Great insights. And Bill, back to you — is that message resonating with customers? We keep joking, but it’s not really joking, because I see more advertising, especially of AIOps, building communities and trying to really explain what it means, especially taking it out of engineering jargon. If I’ve learned anything, it’s that the more organizations progress with cloud, the more we’re moving towards decision-making in a business context, and the executive in a business context. So there’s a lot of education that needs to be done. To cut a long question short: is that message resonating, and from your point of view, what are the key elements we should double-click on to almost cut through the market noise?
Bill Lobig — Vice President of Product, IBM Automation[16:01]
Yeah, I’m smiling as you ask that, because I think Matt sells himself a little short. He runs the IBM CIO’s infrastructure for all of the applications, and IBM is a big company with hundreds of thousands of employees. I meet a lot of clients and I talk about the outcomes we can deliver and the value we provide, and they say, well, are you using this stuff internally? It’s like, call Matt — yes, we are. But to your point, look, we’re focused on making it real. There’s a lot of buzz, a lot of hype — AI, it’s like if you don’t say it, you’re somehow legacy. What does it really mean? And it is resonating. In fact, we have more clients driving successful outcomes with this than Matt, although we love Matt. I’ll give you a quick example: Carhartt, this is a US clothing company — they make clothing for industrial jobs, like people who work on trains and engineers and factories — and they’re seeing 15% spend reduction and 45% optimization in their public cloud. And how are they doing that? They’re better able to handle spikes in demand over holidays, Black Friday — they sell clothing, they have discounts — and this is all done through the mechanisms that Matt described: optimizing their VMs, their Java heaps. Matt touched on performance engineering; I actually started my career in this area. As we used to say, there’s always a bottleneck — it’s just, when do you reach the point of diminishing returns in chasing it? But at the end of the day, no matter how much time you spend sizing and doing capacity planning for your applications, it’s still a guess. It might be a really well-educated guess, but it’s still a guess, because you don’t really know how that workload’s going to spike and ebb and flow. And so this is where technology can help make those determinations and allocations, and drive that flexibility and fluidity of hybrid cloud in real time.
Tom Reuner — HFS Research[18:06]
That’s great. And you’re always giving us a new marketing campaign. It’s not Better Call Saul, it’s Better Call Matt — so we have that one already sorted. Matt, slightly challenging you with the question, or innuendo, to close us out: for organizations listening to us on the buy side, what’s a good point to start? What’s a good point to get your head around this? Where would you start the discussion? I think we’d all agree it’s a complex journey, a challenging journey, but we have to be clear — and I think it’s more implicit, but we know the benefits, and as both of you were alluding to, once you have clarity to know what the goal is, that’s why you have to get on the journey. Having said all that, what’s a good, very practical point to start? Where are you starting this discussion with your clients? Just to close out for today.
Matt Lyteson — Lead, CIO Hybrid Cloud Platforms, IBM[18:56]
Yeah, for me, I would start with some area of your environment where there’s some opportunity, and say, OK, I’m going to get observability — I want to be able to see what’s going on with my containerized applications running on OpenShift — and then I want to bring in something like Turbonomic to make sure they’re getting the right resource allocation. You start with a small segment of your environment, and then you’re able to quickly say, oh, this does work, this is interesting, prove out some of the value really quickly, get the ROI, and understand where you’re going to have those human hurdles. Because one of the interesting things for me: I was egging my team on and I said, go and turn it on, let’s turn it on at scale, and they’re like, whoa, whoa, I’m not sure we want to do this. So there’s a little bit of a mind shift when we say, no, I want the talent on my team, these experts, to not focus on this minutiae and the mundane things — I need them to focus on 23 steps down the road. There’s a bit of a mental shift, and understanding what the dynamics look like with your specific team is an important step on the journey. So start out with something tractable, work through it pretty quickly, and then you get to a point where, OK, I’m going to do next, next, next, next.
Tom Reuner — HFS Research[20:11]
That’s great. I think it’s a great way to close the discussion out. First, I promised the audience outstanding automation pros, and I think you delivered in spades, so thank you for that. But again, there’s so much we can do today — take it as food for thought, reach out to both Bill and Matt to discuss it further. On our side, the HFS side, everything we discussed is almost the next step, the evolution of automation. There’s so much work to be done. As I keep tiring of saying, we have to move beyond this myopic view of whatever the tool is. It’s really about technology, it’s about outcomes, it’s about customer centricity. There’s so much to discuss. So, to all three of us, a first big thank you, but we look forward to engaging with you in further discussion on that. Talk soon.
Bill Lobig — Vice President of Product, IBM Automation[20:37]
Thanks. Thanks, Tom.
Matt Lyteson — Lead, CIO Hybrid Cloud Platforms, IBM[20:58]
Thanks so much.
