As enterprises push AI deeper into their operations, the challenge is no longer simply deploying intelligent agents. It's making sure those agents can work together safely, access the right information and deliver measurable business results. How can organizations build an AI operating model that connects people, data and infrastructure while keeping costs and risks under control?
Tyler Lynch, outbound chief technology officer for the IBM Automation Portfolio at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating Model 2026 interview series to discuss how intelligent automation can help enterprises accelerate operations and turn AI investments into business value.
The conversation covers the importance of closing skills gaps, establishing shared context across AI agents and infrastructure and using knowledge graphs to connect information across enterprise systems. Lynch also addresses agent safety, sandboxing, observability and the growing role of FinOps and token telemetry in managing AI costs. With businesses under pressure to demonstrate returns on AI spending, the discussion highlights why operational visibility, stronger coordination across systems and faster business execution matter more than token costs alone.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM's AI Operating Model Series. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for IBM's AI Operating Model Series
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for IBM's AI Operating Model Series.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM's AI Operating Model Series. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to IBM's AI Operating Model Series
Please sign in with LinkedIn to continue to IBM's AI Operating Model Series. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Tyler Lynch, IBM
As enterprises push AI deeper into their operations, the challenge is no longer simply deploying intelligent agents. It's making sure those agents can work together safely, access the right information and deliver measurable business results. How can organizations build an AI operating model that connects people, data and infrastructure while keeping costs and risks under control?
Tyler Lynch, outbound chief technology officer for the IBM Automation Portfolio at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating Model 2026 interview series to discuss how intelligent automation can help enterprises accelerate operations and turn AI investments into business value.
The conversation covers the importance of closing skills gaps, establishing shared context across AI agents and infrastructure and using knowledge graphs to connect information across enterprise systems. Lynch also addresses agent safety, sandboxing, observability and the growing role of FinOps and token telemetry in managing AI costs. With businesses under pressure to demonstrate returns on AI spending, the discussion highlights why operational visibility, stronger coordination across systems and faster business execution matter more than token costs alone.
As enterprises push AI deeper into their operations, the challenge is no longer simply deploying intelligent agents. It's making sure those agents can work together safely, access the right information and deliver measurable business results. How can organizations build an AI operating model that connects people, data and infrastructure while keeping costs and risks under control?
Tyler Lynch, outbound chief technology officer for the IBM Automation Portfolio at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating ...Read more
exploreKeep Exploring
What was the motivation and thesis behind creating the new AI operating system?add
What key operational variables should organizations prioritize and control when setting up AI-driven infrastructure and governance, and which concerns are most important as they build that architecture to run their business?add
What distinctions should the industry make between deployed autonomous/delegated agents and local "co‑worker" agents, and how can we apply ITDR‑style telemetry, sandboxing, and policy controls to ensure AI agent safety and financial risk management (for example, by revoking an agent's API key)?add
What should organizations measure to quantify the value and impact of AI agents and intelligent operations?add
How should organizations (and individuals) approach delivering customer-focused capabilities and outcomes, and how does your team engage with customers (including the roles of consulting, field CTOs, automation, multidisciplinary work, and the pace of change)?add
>> Hello, I'm John Furrier, host of theCUBE with my co-host Dave Vellante. We're here for CUBE Conversations at IBM's flagship office at 1 Madison in New York City. We just had our Analyst Day, getting all the briefings from all the top executives and find out what's going on in the changing landscape of AI, AI software, AI infrastructure. Tyler Lynch is here, he's the CTO of IBM Automation, one of the key pieces of the new AI operating model. Tyler, great to see you, thanks for coming in. I know you're super busy.
Tyler Lynch
>> Thank you for having me on, it's an honor to be
John Furrier
>> here.
John Furrier
>> I was really blown away because the operating model kind of is consistent with some of the things we've been talking about. The operating system, what's going on at the semis, how these systems are being redefined. But now when you go up the stack and say, okay, the data needs to feed the infrastructure, the intelligence interaction, it creates a huge operating model disruption and acceleration, maybe not a disruption, maybe an acceleration to value creation, which you guys are pointing out. What's your biggest opinion on how companies should think? Because they've lived through the cloud era, They saw the lift and shift, they saw things go to the cloud. Now you have this AI era that is different, always on. You have a lot of things going on with agentic. What's your opinion on how customers should think about framing the operations?
Tyler Lynch
>> So I love the cloud analogy here. When I think about the cloud, the cloud gave you these primitives to build upon and to start to wire together. Well, that's what we're doing with AI, but we end up in the same kind of traps that we fell into with the cloud where we have a database expert, we have a streaming expert, we have this expert and this expert. but they're not really tied together to really realize that value chain is hard. So we took that same problem set and we took a step back and we said, what are the primitives that our customers are working through to generate value today? And AI starts with data, data realization, but we quickly move into these other places of intelligence and around. And so how do we layer on, how do we provide that shared context to really provide value? And that was our thesis behind this new AI operating system.
John Furrier
>> It's interesting you mentioned cloud 'cause I love the example. I even heard the word shift left a few times.
Tyler Lynch
>> You did.
John Furrier
>> And with AI. So you're having a lot of these kind of large-scale infrastructure conversations, shift left, governance is not just compliance, it's a whole nother level. What's the key operational variables that people should have their hands on the knobs? If you had to pick a few, what should people be most concerned about as they set up this architecture to run their business?
Tyler Lynch
>> I would say it starts with people and skills. There's clearly a skills gap that has not caught up to the demand of what our boards and the market is asking our clients to achieve. And so just being honest about what skills gap exists today and say, do I have the people with the skills to achieve the outcomes I'm trying? That's where we should all be starting. And the answer's gonna be no, 'cause a lot of what we're doing is new and emerging and novel, so we can't know everything yet. And so how do we do enablement properly? How do we take these tools that we've planned to use and bring them together? But that really starts with enablement and knowledge.
John Furrier
>> You know, when you have agents now out there running around, they work while we sleep, that applications would just be hosted. There's a whole category of performance monitoring of applications, but applications aren't agents. They're doing a lot of other things. How is that playing out in your mind too? How should enterprise, because we're hearing again, observability, tracing, some of these same concepts are now moving into these agents that are out there, but they're not the same as an application.
Tyler Lynch
>> They're not. I think there's actually another distinction that we need to make as an industry that people aren't talking about. We have agents that are deployed that are operating like applications, and some are autonomous, some are not, some are delegated. But then we have local agents, what I would consider co-work agents. This could be your coding agent. This could be your Muse to be what's in vogue right now, or Dots. Those are very different modalities from authorization, logging, telemetry. And no one's nailed it yet. One of the things that we're looking at doing as part of the FinOps Foundation and the new AI Security Alliance, part of NVIDIA Open Agent Safety Platform, is we're looking at safety for AI agents, but that needs to be extended also to financial safety. So I was on the phone last night with Bill. He runs our Apptio product group. It was on Thursday night, 11 o'clock at night. We're on a Zoom and we're talking about what they're doing in FinOps Foundation. And we had this aha moment. That we can sandbox and isolate agents that are making bad behaviors. Our ITDR detects it. Why can't we do the same thing for financial risk and financial safety? We can now start to emit telemetry directly out of these agents for token usage in and out. Why can't we have policies evaluate that just like we do for ITDR and maybe revoke an agent's API key so it can stop, put it in the corner until a human can go look at it?
John Furrier
>> It's funny, the FinOps event this year, we streamed live and you can see the program got disrupted a little bit by tokenomics because at that time token maxing hit a couple of weeks before. But the cloud had different economics. What's different in cloud economics and say AI economics is now you have faster time to value with the path of inference, for instance. So how do you see that tying together or should they be tied together? What's the frame or mental model that you're looking at when you look at the FinOps? Equation?
Tyler Lynch
>> So I think we're a little different than a lot of other FinOps tools and companies. We look at the complete value chain from the cost of your endpoints, your laptops, your desktop depreciation schedules, your workforce, all the way through your cloud spend. And now agents and tokenization, token maxing is a thing. Certainly we talk about value maxing. Neil, the head of our AI, talks about how do I value max? And for us it is taking all of these signals together and tying back to a true ROI. We have a project that is using AI How do we know we're receiving value out of it? We had a team inside IBM that was spending about $300,000 a month on tokens. Is that a bad thing?
John Furrier
>> Depends on what they built.
John Furrier
>> Maybe.
John Furrier
>> What they built. that's the question.
Dave Vellante
>> But depends on the value.
Tyler Lynch
>> Depends on the value. Well, so here's how I look at it from a CTO lens.
John Furrier
>> There were sales guys just importing PDFs. Maybe not so
Tyler Lynch
>> much.
Tyler Lynch
>> This was all engineering. This is all engineering and definitely engineering product
John Furrier
>> design.
John Furrier
>> But it depends what you build. If you have a paintbrush and you have a canvas,
Dave Vellante
>> But cost per token's not the right metric. Cost per token is not the right metric. It's all about the outcome.
Tyler Lynch
>> So what I did is I have one of my agents every Monday, it goes out and it looks at 250, no, 2,015 repositories and it builds me a dashboard of what my teams are building. And it also bumps that up against their roadmaps. And so I looked at what they're building, their velocity and their roadmaps and how they're knocking things down. And I looked at the cost and I was like,
John Furrier
>> yeah, I can live with that.
Tyler Lynch
>> I can live with that, yeah.
Dave Vellante
>> So, okay, dashboard, you just built the dashboard, you sort of made it happen. What's going to happen to this $50 billion BI business, right? Static dashboards were really useful, but they're not so useful anymore. So how do you see that playing out?
Tyler Lynch
>> I think we saw this early on when you saw people like Amazon QuickSight Q come out where you could start to ask natural language queries. I think Sigma Computing was very interesting because they gave you an Excel-based interface. And people know how to use Excel. Yeah, I think the market is radically shifting away from static dashboards, Tableau, to where every knowledge worker is going to be able to ask their personal agent, hey, what's my pipeline look like? Or what's my product roadmap delivery look like? And those agents will dynamically go create dashboards and SQL queries and generate the information that's unique to me. That's where I think we're going in the next 24 months. Yes, I think it'll be very quick.
Dave Vellante
>> Sigma's interesting. Hex is another one that I think— I don't know if you partner with them as well, but Let's talk about context. It's a hot buzzword. Let's talk about what it means. I frame it this way. You got a lot of operational data. Fine. It's in databases. Everybody understands it pretty well. It's that tacit knowledge that we struggle. We're pretty good as an industry at what happened. We're okay at why it happened. Okay. We're really not good at what's going to happen next and what I should do next. And that's that sort of tacit knowledge that you all and others talk about.
Tyler Lynch
>> I would argue some industries are really good at what's going to happen next, what we should do next. We think about predictive model forecasting models being used inside financial services and healthcare, or maybe pick a random company. My assumption is CVS clearly knows what's going to happen for medical consumption. But generally speaking, knowledge workers at the individual, they don't have that ability, right? It's really an enterprise-grade function and there's a team that does forecasting and modeling. So I think that context is going to be personalized. Probably first at the team level, right, is how we're gonna look at this. And then it's gonna move very, very rapidly. And I think we're gonna move away from just all LLMs for everything to the new decisioning models. You saw Jev come out in the last couple weeks, Amazon really announced theirs last week. And I think we're gonna layer back on forecasting models and predictive models that we kind of forgot about. They're not in vogue anymore. I think we're gonna see a context that is LLMs, SLMs, decisioning models, and risk models and forecasting models. And that's going to be our
John Furrier
>> shared—
John Furrier
>> It's almost like a reset, almost a reformation, bringing back this, and this comes back over the top. I want you to talk about shared context. Okay. That's an important topic because now we're sharing context because if I'm doing this, you might do that. And also talk about the tool problem or tool consolidation or abstraction because we're seeing a lot of tool calls.
Tyler Lynch
>> Are we talking about agents specifically?
John Furrier
>> Agents specifically, yeah. Shared context for the data and then the tool calling. I don't know, it's kind of Wild West. I think that's the word.
Tyler Lynch
>> Oh, it is the Wild West.
John Furrier
>> I got my apps here. Just downloaded Instinct. Jeb, I'm not trying, I'm not using Muse. I'm not going to try that one. But you can see what they're doing. But that's a tool to me.
Tyler Lynch
>> It is a tool. They're all tools. We're pulling them all together.
John Furrier
>> So how they can be useful, how do you rein that in? How do you think about— because you have shared data there. That's good context. Tools, tooling.
Tyler Lynch
>> Well, so shared context, that term is actually overloaded pretty heavily. When we talk about agents and agent swarms and what they know about your business, business process, we see people taking Markdown, whether it be Obsidian or something else, putting them on shared drives, and their agent swarms can use that for definition of what context am I working on, things I need to know, things that are maybe mutable but not often mutable. And so as an example, My source code lives in this repository and Tyler is the owner. Those things are mutable but not highly volatile. Then there's another layer of context graph that is ever-changing with a high volatility of change. And this is what we're building with Infragraph. Infragraph is a graph-based representation of your infrastructure, but we're extending that to be more of a knowledge graph. And so it'll actually be used by our FinOps tools, our observability tools, and they will also enrich that graph. Well, if I can know that this VPC has this EC2 that calls this database, but this EC2 has 65% CPU consumption, as an example, that real-time decisioning can be a game changer.
John Furrier
>> Now, all those graphs are great. Again, like I said in my question, Palantir makes ontology sound like it's new. It could have been around for a while, but if you have an Infragraph, that's cool, but are there layers? Because now there's more relationships beyond the infrastructure. What do you see as the key connection point and how should graphs behave?
Tyler Lynch
>> Why, here's the way I think we're going to end up. We're going to see a bunch of different graphs and they're going to have different contexts and we're going to need some sort of correlation ID to work between them. And this gets to your tool conversation. Well, how do I now access your graph and how do I understand the semantic relationships in that graph, how they relate to whatever's in my graph? And that's the hard part that I don't think we've figured out yet is when Amazon launches their graph and Azure shared Cloud Graph and we have our graph and everybody else. We need the agents to reason over these through tools, most likely MCPs or other things.
John Furrier
>> What'd you call the term, the ID between graphs?
Tyler Lynch
>> Correlation.
John Furrier
>> Is that like a delegation graph or is that how do you, what is it?
Tyler Lynch
>> It is just some sort of ID that I can query between two different disparate data sources and join them together.
John Furrier
>> Okay, cool.
Tyler Lynch
>> Think of it like a relational ID.
John Furrier
>> Okay.
Dave Vellante
>> So knowledge graphs, of course, the appeal is that they're expressive, but historically in order to query the knowledge graph, you had to go back 15 years and in terms of the way in which you interact with it, everybody's figured out, wow, I can now interact with knowledge graphs. It's very simple. And so they're popping up everywhere. So that part, we figured it out. It's pretty easy actually to get a knowledge graph out there. The hard part, as you're saying, is that glue to hang them together and actually interpret the
Tyler Lynch
>> context.
Tyler Lynch
>> And the modeling of the graph. Graph modeling is very different. Remember when everybody tried to move to DynamoDB from SQL and all the projects failed, right? They
Dave Vellante
>> almost—>> They scaled, but they didn't fail.
Tyler Lynch
>> They failed, but they failed. We're going to see the same exact evolution, I think, with context graphs. Just graphs in general, GraphQL, and these things are different. The way you model nodes and edges is not the same way you model a table and rows and columns.
John Furrier
>> We still use DynamoDB 'cause we had nothing else to
Tyler Lynch
>> start with.
Tyler Lynch
>> I love Dynamo, I've worked at AWS for years.
John Furrier
>> We had nothing else to compare it to, so it's great. When you look at the intelligence piece of it, because now you have the data piece, this operating model that you guys were presenting is, I call it the data layer, it's my word, but it's really more than that, but the underlying semiconductor movement is going to get much enabling more fast speed. What changes as that intelligence moves up the stack? It might get cheaper, but it's not going to change the value. That's what came out of the talks today was clearly that value won't drop. Prices might open weights, might introduce new things. Commodity models. Rob Thomas, he thinks they're going to be commoditized, frontiers commoditized, but not on the value, but on price maybe, or differentiation. What's your view on that? Because as the horsepower and the supercomputers get better, we've heard the Z's got a roadmap till 2045 from Arvind. So there's a lot of gear out there, big iron, and you get the cloud, NeoCloud. So they got a lot more coming.
Tyler Lynch
>> Well, so I think that we're all looking at the energy crisis as something that's real. I believe Amazon is the largest power buyer in the world. And so the reason they adopted their Graviton chipsets was because they could save essentially 90% on power over x86. We're also looking at the same thing. Our Z-Boxes are generally 20% more power efficient, 20 times more power efficient than x86. So I look at, think about where's the value? It's going to be power consumption. Number one, models, token costs are also going to change. Everything gets cheaper as you get commoditized. Yes, but the value is not going to change. What I really question though is how much value are we getting out of these? Yeah, my husband works in financial services and the conversations we have over dinner, oftentimes the lead product is how do they bring agents to this? Well, what value are you getting out of it? What's the capability you need? What do you want to focus on?
Dave Vellante
>> So what is he measuring? You're talking about Arvind said we were 3% of budgets, now we're 10% of budgets. We're probably going to go to 20% of budgets. There's going to be value there because we're basically scaling with less labor. What should people be measuring?
Tyler Lynch
>> So for me, speed of business matters, and that's really hard to quantify sometimes. And so I talk to people, can we get something to market faster? That's not just software development, that is maybe an underwriting process. Can we get through an underwriting process instead of 2 or 4 days? Can we get to it in 20 minutes? Underwriting is a very expensive process for mortgage origination. And so I'm looking at things that aren't so easy to quantify for value. It's not just did I fulfill an order. It is did I do it faster? Did I do it better? Do I set myself apart from my brand and from my competitors?
Dave Vellante
>> But ultimately you can quantify that in productivity.
John Furrier
>> Your scenario actually is a good one because a lot of smart people, their first reaction is saying let's throw some agents at the problem. Well, what problem are they trying to solve? Or hey, agents could help automate my job because it's a natural instinct to say I do a lot of Undifferentiated heavy lifting or toil. Don't we all? They always say, "OK, sure, I'd love an agent." Who doesn't want agents? throw me an agent. But talk about— it's not just throwing over the transom. That stuff has to get locked in. This is the number one conversation we have on theCUBE, which is, yeah, it's easier said than done to do
Tyler Lynch
>> agents.
Tyler Lynch
>> Well, do you trust agents to do everything for
John Furrier
>> you?
John Furrier
>> Oh, God,
Tyler Lynch
>> no.
Tyler Lynch
>> I don't have them write my emails. I don't have them write my blogs. I don't have them write my white papers. I write those originally. Now, I'll have them edit it and do Socratic methods and things like that to make me a better writer. Or better technologist, but I'm generally not having them be the originator of thought.
Dave Vellante
>> But you do have them as sort of, help my buddy. Help my buddy.
John Furrier
>> Yeah, your buddy, Zach.
Tyler Lynch
>> My buddy.
John Furrier
>> All right, intelligent ops. What does that mean? Because I love that term. It was one of the new releases, a flagship launch for IBM. What does that mean in the future? If you had to go put the magic wand and
Tyler Lynch
>> say—
Tyler Lynch
>> So let's talk about intelligent operations, but it starts with operations today. So when we talk about infrastructure operations, we're making changes to infrastructure based on needs or availability, security. If we can make those intelligently with machines, we can start to do very interesting things. So as an example, we were building these intent-driven workflows, conversational interfaces for mainframes, and they said, we'll go use this storage address and apply a new LUN. Well, they did it. Unfortunately, that storage address was actually in use by somebody else and we deleted all their data. I'm going to restore their data.
John Furrier
>> I've heard many stories.
Dave Vellante
>> Hopefully it was backed up.
Tyler Lynch
>> So, but what Intelligent Operations gave us is, well, now before we make a change, we can go verify that that address space is not in use anywhere else. Now, a human might have done this, but they might not have done it, right? They were given an address, the data guy goes, oh, okay.
Dave Vellante
>> So they might have done it Monday, but not the year later. Right, right.
John Furrier
>> I'm going home and get it done and up and then look. Oops, I forgot.
Tyler Lynch
>> So for us, Intelligent Operations is taking a human's intent or signals through telemetry, monitoring, observability, FinOps signals and taking some sort of action. So we're doing this with our Turbonomic product where we know that you are oversizing your EC2s for general usage. Let's downsize them for you. That's pretty easy. That's a standard one. But I like the more complex use cases, which is catching the errors that humans might not catch. And so is this thing in use? Dependency impact analysis is something I'm very much interested in as a scientific area where if I have Terraform that manages easy resources, but those resources are actually being used by 40 other consumers, This Terraform doesn't know about it. Can we do dependency impact analysis to determine would I adversely affect anybody else?
John Furrier
>> Yeah.
Tyler Lynch
>> And the person who owns this software wouldn't know about it. That's where I think it's knowledge graph to do that. We can use a knowledge graph to do that. Well, we're looking at it right now.
John Furrier
>> That highlights a couple of things that are going on now. Disaggregated serving inside big AI factories have a lot of resources that need to be known. Distributed databases, another hot area.
Tyler Lynch
>> Yep.
John Furrier
>> A lot of dependencies. So we're seeing more complexity coming if that continues.
Tyler Lynch
>> We are. And I think intelligent operations will also bring scientific rigor where it doesn't always exist. And I'll quantify, we have coders and we have software engineers, and then we have people that manage infrastructure and you have SREs. These are very different in discipline. The coding agents or agents can actually act as scientists and will know things about dependency impact analysis or CAP theorem. Consistency, availability, partition tolerance. The agents know this stuff, but not everybody that's managing your Cassandra database or your DataStax database or your Dynamo does, as an example. All
John Furrier
>> right, IntelligentOps. All right, we got the IntelligentOps. Now on Apptio side, what came up in the talk was, I wrote it down, I think these are my notes, it may not be verbatim. Apptio is a maturity issue. Maybe that's my notes, via Cloudability and Kubecost. So now you have maturity was thrown in there, and so we're seeing the maturity still growing. There's a lot more maturing to go. How should people think about the maturity curve that for AI is different? I
Tyler Lynch
>> think if you frame your note a little bit, what I was talking about is we think about this in a phrase or a phased maturity for people that would look at FinOps. First, they need to observe what they're working on, and that's generally retrospective, last day, last month, last week, and start to make changes. Generally it was a month back, right? We got our bills from our cloud. We're like, oh no, let's go change something a month later. Well, it takes another month to make that change.
John Furrier
>> Yeah, secure that S3 bucket.
Tyler Lynch
>> And then we're moving past that. Now we're into, we're doing real optimizations in shorter timeframes, same day, next day. But that information is actually being fed back into our CFO and our controllers where they can start to plan and budget and use forecasting based on real data, real spend, and it matches our cycles. And that's what I think.
John Furrier
>> So there's measurement in place, there's maturity, it's just natural things happening. But you guys are like, okay, I'll re— it's not an issue. It's a maturity concept.
Tyler Lynch
>> It's a maturity concept. And this is how we deliver incremental value to our customers. They have to start somewhere. Our software is commercial-grade software and it often has many features and capabilities. And this is true for lots of other software, but you got to figure out where you start and provide the most value and how you start to bring that up to outside other parts of— it's not your DevOps team, now it's your FinOps team, now it is your controller, now it's your CFO. How do we inform them all? And you can't do that all on day one. We land and then we expand.
John Furrier
>> Yeah. Well, I really appreciate you coming on theCUBE, but also talking to the top industry analysts and giving them kind of a peek into what's happening and what's coming out. You got TechXchange coming up, which is going to be a customer and technology exchange going on, peer review, peer sharing, IBM Think. What message do you want to send to the market around what this platform is, this operating model, and how should people think about it? And ride that wave?
Tyler Lynch
>> I think that we all need to think a bit broader outside of our normal day-to-day scope. First off, instead of just being a developer or whatever that is, take a step back and think about capabilities and outcomes that you want to deliver for your customers and your clients. That's where we should all be starting. It's going to have multiple pillars, multidisciplinary, and that's what we should all be doing. When it comes to your engagement with us, we can do that through our consulting arm, You could also reach out to me or my field CTO team. We meet with customers a lot to just have conversations and it's free of charge and kind of reshape that conversation on outcomes and capabilities so we can tie that back to ROI.
John Furrier
>> And automation is a big part of the operating model. And you get to that center of that Venn diagram, which I love. You got to do a lot more stuff. The more capabilities come in slowly or fast, coming in hot.
Tyler Lynch
>> We're trying to be intentional and we don't want to move at the speed of some other companies. We want to be intentional. People come to IBM because they trust us. And we bring rigor and trust, and that is our product strategy and the way we deliver for our customers. So we're intentionally moving a little bit slower. I wouldn't say it's a lot slower. We're doing a lot. I worked a lot when I worked at Amazon. I am working a ton now. I feel like the velocity is super
John Furrier
>> high.
John Furrier
>> Yeah, you guys are doing some good work. It's operating system of trust too. That's the outcome. Tyler, thank you for coming on theCUBE. Really appreciate it. Thank you, bro. I'm John Furrier with Dave Vellante we're here from the IBM One Madison, beautiful flagship building in the heart of New York City. Thanks for watching.
John Furrier
Dave Vellante
