1. Shaping Cisco's Portfolio Strategy with Jeetu Patel and Jeff Schultz

Jeff Schultz joins Jeetu Patel to delve into the evolution of the Cisco Portfolio and how this focus on critical business outcomes for the AI era is delivering real results for the ecosystem.

Transcript

SPEAKER_00 0:00

Hello, everyone. Welcome to Cisco's Inside Product Podcast. And today I have with me Jeff Schultz, who actually runs our portfolio strategy. Welcome, Jeff. It's great to have you here. Great to be here, G2. So you have been the kind of mastermind behind the scenes on like helping us put together the cogent, coherent strategy around the entire portfolio, rather than talking about every different different piece part. Talk to us about the overall like how how should customers process what Cisco is doing and why it's so valuable.

SPEAKER_02 0:36

Yeah. So I think we traditionally have gone to market as a whole bunch of different products and different product lines. And what we really wanted to do was focus on a couple of the key outcomes that our customers care most about, especially as we move into this next era of AI, which is agentic AI and physical AI. And so we you can really look at what we do across networking and security and observability and collaboration and break it down into a couple key outcomes. So AI ready data centers, because we know our customers are looking at how they rethink their data center, rethink what they do within the data center and across data centers, where you know, future-proof workplaces, so how they think about where humans work and connect with customers and build things, how that is fundamentally changing in the world of AI. And then, you know, foundationally digital resilience. So how do they think about leveraging data to make sure that anything that can potentially bring down their business, whether it's an infrastructure problem or a security threat or agentic behaviors, all of these are looked at at scale and at machine speed, and we can help keep the business up and operating?

SPEAKER_00 1:41

So if I were to take a step back, what you're saying is that the the key outcomes that we focused on are three. Yes. For a company that's our size, building AI ready data centers, making sure that our workplaces for our customers get to be feature-proofed, and making sure that there's a level of resilience in the infrastructure that we call digital resilience. Yes. So given that, um let's drill into each one of those. Uh how should you know, why is Cisco different in each one of those areas? And specifically, what are the capabilities that customers can benefit from? And what is the problem that we're trying to solve for those customers in those domains?

SPEAKER_02 2:21

Yeah, so start with data centers. So in the data center world, what we know is there's a variety of different kinds of companies that are building data centers. So you have hyperscalers, neo clouds, service providers that connect all of these data centers, enterprise customers. And ultimately what they're all looking to do is how do we figure out how to scale these data centers in the world of agentic AI? And um and at the same time, though, how do we make sure everything is secure in this new world where everything is hyperdistributed? And how do we make it easy for them to adopt this technology? And so there's really those are really the three key things that we're looking at within the data center realm. So how do we help them scale AI? You know, we do that through um, you know, silicon and systems and optics that help them scale out within their data centers, scale across data centers and act as if multiple distributed data centers are really working as one, um, fusing security into the network, so a lot of innovation around how we think about security and networking becoming one. So our network devices literally become security devices as opposed to having to have a separate set of technology in the perimeter for security. Um and then um full stack AI. So how do we bring together our capabilities as well as our partner capabilities? You know, this is where we do a lot of work with NVIDIA on the secure AI factory to tie together um all the technology and make it easy to package for our customers to uh deploy.

SPEAKER_00 3:47

And so when we think about um the fact that we have a multitude of different customers from hyperscalers to neo clouds to sovereign clouds to service providers to enterprise, um does um do these areas work across all of them? Are we doing things that are specific for each one of these areas? How how do you think about the entire kind of value proposition of Cisco?

SPEAKER_02 4:11

Well, I think what's what's really cool is that this you have to think of this as a co-designed stack, a vertically integrated stack. So it starts with silicon one, but it uh it goes all the way from sort of the silicon to the semantics. So the silic silicon, the systems that we're able to build on that, the software, the models, everything that we're able to do is co-designed in the sense that um as we think about um customers, none of these things are working in silos, right? So our enterprise customers are leveraging service provider technology, they're connecting to hyperscalers and neo clouds. And so the goal here is um by doing this unification, we are able to really dramatically simplify how our customers think about deploying an architecture that expans all of these different areas and an architecture that leverages the power of the stack. So, really thinking deeply about what we can do in silicon to enable features within our systems and within the operational side of the house that um you know that all leverage each other as opposed to working independently. And by the way, all of this being done with a mindset of open ecosystem, because we also know that we're not the only vendor that that is um involved in these solutions. And so um part of our design mentality is how do we build this so that you know you have a compounding value or a compounding effect as you start deploying Cisco technology, but at the same time it works with all of the other technology that our customers have in these areas.

SPEAKER_00 5:34

So I I want to dig into a couple of those areas, but before I do that, uh you and I have worked together for God knows how long, 15 years or something. Yeah. Um and so when you came to Cisco, um tell me what surprised you on the positive and what surprised you on the negative, and what have you tried to do on the thing that surprised you on the negative to make it better?

SPEAKER_02 5:57

Yeah, so I mean what well let me start with what the positive. And I I remember the moment that I I mean I it was it was very, very, very cool because I I think we were in the still in the time of COVID and and the pandemic, and um we were very much looking to-exactly the location where I called you to say, I think you need to join here.

SPEAKER_00 6:16

Yeah, I kind of remember. Yeah and uh I was calling you and like I'm like you just need to come here.

SPEAKER_02 6:24

Yeah, yeah. And I think and I think the conversation was about the fact that I think we you we were very focused on the collaboration technology, and I was very passionate about wait a second, you're gonna have to bring collaboration and networking together to be able to address the real problems that our that our customers are facing with with you know work from home and hybrid work and things like that. And so um, you know, so the the the positive was you look at the portfolio and it's incredible, like leadership areas across everything, you know, networking, security, observability, um data with data with Splunk, um, the you know, collaboration technology, device, both the devices as well as the software side, yeah. Incredible assets. Um the negative was that they were all operating in in silos. And and I think that um, you know, that hybrid work you know effort was really our first attempt at saying, wait, we really need to think about breaking down the barriers between these organizations and talk about how they how you know one plus one is greater than two. And um, you know, and I think we we successfully did that. I think then one, you know, this whole the next phase of what we've done as we've really brought together the rest of the portfolio is thinking even beyond that. So, you know, it started with hybrid work, now it's agentic AI, physical AI, and it's another example of the fact that I don't think you can um address these challenges in silos. That you really need to think about how all of this comes together.

SPEAKER_00 7:43

Why is that good for customers? Why do customers care about that?

SPEAKER_02 7:46

Um I think you know what's happening with you know, there's so many things happening with agentic AI in terms of the opportunity that you have, as well as the the you know, the potential dangers, um, both in terms of you know the attack vector that increases, um, the the you know, what could possibly go wrong if an agent goes off outside of its guardrails. And um this is all happening at machine speed and scale. And so that has dramatically changed how our customers have to think about um you know the technology ecosystem and their infrastructure for dealing with this. So, like for instance, you can no longer count on security being a separate stack from from your network, in our opinion. I mean, you you need to start thinking about how do you inspect traffic at line speed, how do you how do you use the network and security working together to be able to contain uh a risk or contain an exploit or contain an agent that might be going rogue. And um again, if I it I think that given how AI has fundamentally changed the game in terms of this speed and scale, um you we need to start breaking down the barriers of these products um even more.

SPEAKER_00 8:52

So, okay, so let's you've talked about three or four things. I I want to kind of let's break it down. So platform is the first one.

SPEAKER_01 8:59

Yep.

SPEAKER_00 9:00

Walk everyone through systematically how we think about the platform, uh, and then we'll go from there into open ecosystem, but start with the platform first. Okay, so what is a platform? Because it's an overused term right now in the industry.

SPEAKER_02 9:12

Yep, yeah, yeah. So I think there's a couple different aspects of our of our platform strategy. So first, we already talked about this vertically integrated stack. So being able to leverage the fact that we have silicon, that we have systems, that we have software, and making those all work together in new and innovative ways. Um, other areas that we think about from a platform perspective, fusing security into the network. So really taking our security assets and our and our networking assets and thinking of how they come together as one to leverage, you know, to both be faster as well as leverage the network for things that you really will have to do there and not not in any side um set of applications. Um the third aspect of platform is data. So, you know, being able to bring together data that has traditionally been very siloed in organizations, network data and traffic and understanding of processes that are talking to other processes, tying that together with security information and security threat information, tying that together with the experience data that we have, are people actually connecting and having good experiences across all of these networks that we're building? Um, all of this data, um, being able to bring it together is part of this platform story that I think is truly unique in the industry. And then last but not like the four fourth pillar of the platform is how do we then simplify this using AI and what we call agentic ops? So being able to then create really brand new user interfaces that allow for human and agentic collaboration that tie together all that data and infrastructure we just talked about in unique ways.

SPEAKER_00 10:43

So, okay, so that that makes sense. There's a couple other things that I at least think about on the platform. Tell me if you agree. One is I think if if we do our job correctly, then the marginal cost of ingestion of every new technology should go down dramatically, and we should almost have that curve approach zero, where every new technology we get can be ingested by the customer for a very, very low marginal cost. And number two is there should be a compounding value that's being delivered. So everything that we deliver in the market should be great for not just what it does, but also what it brings to the table for everything else that you have from Cisco. That's right. Do you agree? Yes. And and so now talk about given that, what does the open ecosystem bring to the table?

SPEAKER_02 11:28

Yeah. So let me let me give you an example though of like the compounding effect. So I think I think infusing security into the network, that's a really good example. So the new smart switches, as an example, which now allow you to run firewalling capability on the switch itself. As you and by the way, the switches in in and of themselves are incredibly powerful switches. So as you start bringing them into your environment, what is happening is not only are you driving up the level of performance on the network within your organization, but you're now actually providing more points where you can do micro-segmentation and inspection of traffic at the same time without having to add additional firewalls on top of it. It's all sort of coming in that same box. And so that allows you to then have a single set of policy that governs both how traffic is is happening and routing is happening within your organization, but then also how security policy is being inspected. And again, every network point becomes a security detection and enforcement point. At the same time, what we've done is we've opened that up so that it works with third-party firewalls. So so at the you know, you can now have one place where you're looking at policy for segmentation that you can now enforce in Cisco's firewalls, you can enforce it in Cisco's networking um gear, and you can also enforce it in third-party firewalls. So you have that compounding effect as you add more and more.

SPEAKER_00 12:53

And so this open ecosystem kind of um concept that I think we started this right around five, six years ago, where we said, hey, um you can't be arrogant enough to think that you know customers are gonna always buy just your technology and nothing else. And so we owe it to our customers as good partners to make sure that if they've invested in company A and Company B, we happen to be one of those two companies, but the first company happens to be a competitor of ours, we should not let that get in the way of creating a mechanism to add value to our customers. That's right. Talk about how that's manifested itself and what what are some of the examples of how that's actually come to life.

SPEAKER_02 13:37

Well, let me give you let me give you another example. So I think firewall is a great example of that. But let's talk about the future-proof workplace. And you know, there we talk a lot about modernizing campus and branch infrastructure. We talk about um security for um for both agents as well as humans, and we talk about delivering uh intelligent experiences in that workplace. I think that's an area, again, where um the work that we've done within the collaboration portfolio, specifically within devices, was really focused on open, you know, a couple of years ago we opened that up to enable other vendors to work on our devices, but not just work on those devices. So, you know, think Microsoft Teams, you know, Zoom, Google, um, they can work on those devices, but also because those devices have AI built in, we are able to actually make the experiences of those platforms better because they're running on our devices. And at the same time, because those devices are connected to our network, we're able to gather data, essentially use those devices as sensors on the network to provide our customers with even more data on the experiences that their customers are that their users are experiencing in those workplaces and be able to adjust those workplaces as needs change.

SPEAKER_00 14:50

Yeah, and so this this kind of notion of having an open ecosystem requires two-do tango. Um and so what message do we want to have for our customers as it pertains to the integration of um you know our technologies with our competitors' technologies even? And the open ecosystem at large.

SPEAKER_02 15:10

Yeah, I mean, I think I think you know, you know, like we're gonna be tightly you know, tightly integrated, loosely coupled is the is the term that you use. So I mean we're all we're we're gonna always, of course, look to deliver the great you know, great experiences on our own stuff. But when we talk to our um to other partners in the ecosystem, it's really about how you know we are going to bring out the best of their products and allow them to operate within our infrastructure and add value to that on top of what they would produce or do natively or or with with some of our competitors in the infrastructure side. Um and so I think there's a really good joint value proposition there. Um another area that um, by the way, we haven't really talked about this yet, but another area is on this agentic ops side where we recognize that every vendor in the market is thinking about how to apply agents to help improve um the operation of their of their solutions. You know, what we're doing is um, you know, with with the launch of Cisco Cloud Control, we are talking to our partners about how their agent frameworks can fit into our agent framework so our agents can work with their agents to deliver um solutions uh and and better operations.

SPEAKER_00 16:17

And that's a really good point. Take a step back, and given the fact that you talked about platform, talked about talk about Cisco Cloud Control and specifically what does that do from a value perspective for customers? Like why is that important and why is that not just yet another product release from another vendor?

SPEAKER_02 16:36

Yeah, so you know, I what I what I think is interesting about cloud control is that when we first set on this journey, it ri a lot of it was around simplification within our portfolio. How do we bring together all the different dashboards and controllers that we have within the Cisco portfolio under one roof? And I think it has evolved way beyond that. Um so where we are with that right now is you can think of it as not just unifying our own dashboards, but unifying all of this cross-domain telemetry. So uh networking data, security data, application data as part of observability, um agentic behavioral data, all of this coming together in one place that we can then use, and we can then use that data to apply agentic ops to these problems of how do we deliver better experiences across our infrastructure? How do we constantly look for whether there are performance problems, whether there are configuration issues, whether there are compliance issues, how do we help humans troubleshoot problems much faster than we could ever uh troubleshoot them before, or even give them autonomous capability to be able to keep their networks and their systems up and running. And doing that all in a very open ecosystem where these agents are things that our customers can build and control, but then also, as I said before, work with our partners to have agents working with agents. Because again, if you know, if you have Slack or you have ServiceNow or you have other types of solutions in your environment, you want to be able to have our agents talk to their agents and be able to jointly solve problems. And so this has really become a harness for this agentic operations type which ties together infrastructure, um, security, observability, all in one place.

SPEAKER_00 18:19

Basically, an infrastructure and security harness for organizations that are looking to manage infrastructure at scale uh in a secure manner. That's right. That's right. And so uh as you think about that, how does this notion of agentic ops, which is agents are gonna be ambient and present in your network and they're gonna keep doing what they're doing in the network to monitor the network on an ongoing basis? Walk through what that what happens when an anomaly is detected, and then what do those agents end up doing and why is this agentic ops thing so different from what AI ops used to be like or what the previous generations of these technologies look like?

SPEAKER_02 19:00

Yeah, a lot of it, well, first of all, it starts with the data, you know, having access to more data than we ever had access to before. But I think agents bring a whole new set of reasoning capabilities as well as tool access. So now you have the ability to have these ambient agents um start to reason, start to reason with humans, start to reason with each other, and um really take a lot, do a lot of the steps that we would normally do as humans to sort of understand what a problem, you know, what problem is happening, where it's happening, how to potentially fix it, it can start to do all of that for you. And then you you really, as an organization, you have a choice. How far do you want to go in automation? Do you want to let these agents go ahead and do that and actually solve the problem for you? Or do you want to surface it in new and unique ways for for humans and teams of humans to work with the agents to decide on the on the best solution? And so it's really that reasoning and that tool access that make a huge difference, but data is the key. And and again, that's another area where um if you don't have the data, then you can't work across these domains and you're you're gonna be your agents are gonna be handcuffed.

SPEAKER_00 20:05

So you know, uh, you and I both talk to Onu Rog a lot, and one of the things that he has really articulated well is this notion of I'm gonna have these ambient agents in my network. They're gonna monitor what's going on. If there's a breach, there's an outage, there's some kind of anomaly that gets detected. Based on that, the agent then does deep reasoning and figures out not just the fact that there's a detection of a problem, but what the ideal solution and response and remediation to that problem should be. Yep. Spins up a digital twin. That digital twin could be either emulated with real um you know kind of devices or simulated with an actual uh you know simulation engine. You run live data through that, see how that performs in both an emulated and simulated world. If it performs well, then you can actually apply that patch or apply that change to your system or the policy to your system. And if it doesn't perform well, then you know that it's not ready to be applied. Um as you think about that entire kind of apparatus, um where did you see us um where do you see the customers asking for the most amount of help right now? Like what are the biggest pain points in customers as they think about this infrastructure management at scale? You talk to a lot of customers.

SPEAKER_02 21:22

Yeah, yeah. I mean I think I think you know what what what I think is interesting is so what you just described is really what customers are asking for. So like how do I, in this world where um our environments are becoming more complex, attacks are going to become more frequent, you know, exploits are gonna happen, how do we move faster, and how do we break break down those barriers so that you don't have multiple teams having to use manual processes? So, you know, that that is um critical. But I would I would actually add another thing, and this ties into the other uh sort of pillar of our of our message that we talked about, which is digital resilience. What our customers are also saying is help me get. Control of all of this data because I have more telemetry and data than I've ever had to deal with before. But I need to do it economically. So I need to do it at scale and I need to do it economically. I need to be able to also, even whether I'm creating agents on my own or I'm using your agents, I need to be able to monitor their behavior and trust that they're doing what they're supposed to be doing. So it's great that we can talk about all of this, but we also have to be um conscious of the fact that agents can go off their guardrails. They can be, they're non-deterministic. So our customers are like, give me agentic ops, but also give me a way of monitoring their behavior. And then lastly, how do I use agents even more in the security world? How do I make sure that I have an agentic sock so that you know I know that if something is happening, I can see it quickly, I can respond to it at machine scale. And so that's uh those are some of the core components of our digital resilience pillar that really matter. If you don't have those, you can't have agentic ops.

SPEAKER_00 22:57

And and what has been the child because digital resilience, the you know, the way that we actually deliver it to the market, the the core underlying technology is Splunk.

SPEAKER_01 23:07

Yeah.

SPEAKER_00 23:08

Um what I've heard about Splunk is hey, this is great tech, it's super expensive, and I can't really um go out and justify spending so much money on it because the ingest cost can get to be out of hand because there's petabytes and petabytes of data that need to get ingested. Talk us through the strategy on Splunk and what what is actually happening in that part of the portfolio.

SPEAKER_02 23:30

Yeah, so I mean that's what that's really what we're addressing as we think about what we're what we launched as the Cisco Data Fabric.

SPEAKER_01 23:36

Yep.

SPEAKER_02 23:36

So think of the Cisco data fabric as taking the power of Splunk, but that also contains additional capabilities such as the new Federation capabilities, which allow you to have access to data without necessarily having to ingest it from other sources. So there's one of the big costs is not just ingesting data into Splunk, but it's actually getting it out of the cloud systems and other types of systems where that data exists. The egress costs are super high. So um Cisco Data Fabric addresses that by allowing you to leave data in place and still get value out of it as part of the Trevor Burrus, Jr.

SPEAKER_00 24:08

You're taking your search and analytics to the data rather than bringing the data to Splunk. That's right. That's right. Um that's available today now. That's that's available and AWS, Azure, Snowflake, all of that, right?

SPEAKER_02 24:19

Yes, yeah. And that's a that's an area of massive innovation investment for us. Um so we you know, we are going to, you know, our goal is to allow you to work at data at ludicrous scale. We kind of put a little bit of a joke in there. Yeah. Um but um and and in order to be able to do that, we have to do, we have to make it economical. Um we also have to apply AI to it. So being smart about where data resides, smart about how we how we label data as it comes in. And so there's a lot of effort in the Cisco data fabric around that.

SPEAKER_00 24:46

Aaron Ross Powell, and the kind of data that Splunk specializes in, because data is a broad term. Yeah.

SPEAKER_02 24:50

The kind of data that Splunk specializes in is time series data, logs, metrics, traces, all the data is you know really very highly unstructured that is.

SPEAKER_00 25:02

And gives you a temporal view.

SPEAKER_02 25:03

It gives you a temporal view. So like if you can if you can take this highly unstructured data that has that is time series and start to correlate it along time, you can start to see the the cause and effects of things across your entire environment. Um that that was the you know secret sauce of Splunk from the very beginning, really starting off doing that with like web logs and other types of things long you know back in the day.

SPEAKER_00 25:27

But by the way, interesting story is you actually worked with me at box, but then you went to Splunk for a bit. Yeah, that's right. That's right. And uh ironically enough, they came back to us. So that's right, that's right.

SPEAKER_02 25:36

And this has been a continuous issue, but like agentic AI completely changes the game. I well, actually, one of the things that we haven't really talked about is that every agent action, you think of it as a routing challenge, a trust decision, and a telemetry event. Every single thing an agent does is generating some sort of telemetry that we can use to determine whether it's doing something that's a threat, that's doing something bad, or whether it's it's just a behavioral anomaly, is it just not doing something well? Um, and so we can actually now so that's telemetry, that is time series data. We can now bring all of that together. And so the area of the another area of huge investment within Splunk, after we have the data, is this whole area of agentic observability. So being able to take that telemetry and not just again, not just determine whether it's a threat. Is it deleting as an agent deleting a thousand files? Maybe that's a bad thing, right? We can that's important. But you know, what if you have an agent that's um giving incorrect refunds or or you know, particular talking particularly snarky with customers and creating just a bad, you know, bad experience? Um that is traditionally very, very hard to monitor and understand. And what we're able to do because of the data and because of some of the you know, the Galileo acquisition and some of the other work we've been doing in Splunk Observ Splunk observability is really be able to give developers the ability, the the agent builders, people who are building agents, the ability to test their agents as they're building them, but then also monitor agents at scale to make sure that they're not doing bad things.

SPEAKER_00 27:07

So, Jeff, one area that actually has really started coming up lately is this notion of as agents are proliferating more and more, uh as we're using agents on our daily basis. I know you're a very avid user of agents you know yourself, but the cost of tokens can sometimes get to be prohibitive and almost break the bank.

SPEAKER_01 27:27

Yes.

SPEAKER_00 27:28

And um it's one thing when you have, you know, you break the bank with a couple hundred bucks, it's another thing when you start doing that at scale and you break the bank at hundreds of millions of dollars. And so this notion of governing token economics or what they call tokenomics is a pretty important kind of problem to solve for. Yes. Uh and I think the industry is um is gonna need to collectively solve it. What are we doing at Cisco around that?

SPEAKER_02 27:52

Yeah, so I I mean it is uh one of the most important problems that we have to solve because like none none of the benefits of a gentic AI or physical AI really get paid off in an organization if you can't afford it and you have to start throttling it or overprioritize. And I think that that's the biggest thing that we need to address. And so um I there's a couple of things that we're doing about it. And I think, and and I and I think what's really important to understand is that you need a platform approach to be able to do a good job with tokenomics. You need to be able to see where tokens are being utilized everywhere within your organization. And that means if you are building agents and say you're deploying them in your data centers, you need to be able to be measuring token performance for those agents in the data center. But we also know that like vast majority of agents being deployed within organizations today are actually being deployed um on an employee's laptop or on a desk desk side computer, right? They are Mac Mini and the Mac Studios are actually You can't buy them. Flying off the shelf. You can't even buy them right now. And so um though like you know, so you're running, you know, Claude or Codex or OpenClaw or any of these other systems there. And um and so it creates quite a challenge because it's one thing to be able to monitor token usage within your data center, but if the majority of that token usage is happening outside the data center, you really need a way to look at this um holistically. And by the way, one of the interesting stats we just released that every agentic action r requires 450% more network traffic than a human action of you know, equivalent human action. And that's because of all of the token processing that's happening. And so um, you know, what are we doing about it? Like ours, if you look at what we've done within Splunk and how that now ties into the different parts of our platform, whether it's AI defense or whether it's secure access and things that are running on your desktop, we are looking, we are going to be able to provide that complete holistic view of where tokens are being accessed, whether it's happening again in your data center, whether it's happening in a desktop in your workplace environment. And with when you have that and again bring all that data together, you can now give customers a view of um, you know, is there an agent that is going a little bit rogue in in terms of token processing, maybe too expensive, maybe it's got a bloated skill that it's calling and that's just too too expensive to process. Um, we can now detect that and give our customers the ability to stop that or or put containment around it much more easily. And over time, I would expect this to go into an area where we also give our customers the ability to then prioritize um those actions.

SPEAKER_00 30:27

So being able to look at the tokenomics, which kind of actions are you gonna actually serve and which ones are you not because it's gonna get queued up to be a lower priority? That's right.

SPEAKER_02 30:34

Like a gentic ops is hugely powerful, but you're gonna need to, you know, our customers are going to need to decide what agents are gonna be prioritized over other agents, what problem is going to be a good idea.

SPEAKER_00 30:43

It's kind of like the next level of like, you know, I don't give my daughter um Netflix access in the middle of the day because she might take away my bandwidth if I'm in a bandwidth constrained environment at home and I've actually put that rule in my router and in my switch. That's right. Yeah. Yeah.

SPEAKER_02 30:57

I mean, there's like uh if you think of agentic ops, there's a ton of compliance checking you can do. Or do you want to do compliance checking every single night in in lieu of other problems, or do you want to space that out over time? Those are the types of uh decisions our customers are going to need to make, and we can help them make sure.

SPEAKER_00 31:11

So we are providing a full uh stack of observability for AI that includes um the resilience of the infrastructure, what is your GPU utilization for your agents as an example? It includes guardrails around agent behavior. So is the agent starting to go awry? And if it is, what do you need to do to make sure that you can actually keep them within the guardrails? And third, what are the economics of consumption of these tokens that we need to make sure that we keep in mind because every organization is going to struggle with this? And if an agent is starting to go nuts and start to consume a lot of tokens, you can actually either intercept that agent or kill the agent right away. Aaron Powell That's right.

SPEAKER_02 31:54

And and again, the it's it's all coming together within the Splunk observability platform, but leverages all the rest of the platform to be able to make that happen. And if you didn't do that, you would not have complete visibility.

SPEAKER_00 32:06

So if you were to have a magic wand and you said if a customer used our technologies in these ways, they would get the most amount of return from it. What would you advise customers?

SPEAKER_02 32:19

You know, I think I think that the the my my biggest piece of advice would be, you know, if again, be first of all, I would say don't be scared of the agentic AI movement and physical AI, right? I think these are huge opportunities for us to harness. Um but I think you need to be very thoughtful about how you both leverage it as well as put guardrails around it. Guardrails around it. And um, you know, I think that's why that is why I'm excited about what we're doing to bring the portfolio together to address this, because I don't think you can do that if you look at the problem in individual domains. And so where, you know, the work we're doing to really bring those c those domains together, both from a data perspective, a functionality perspective, the the platform is really critical to our customers being able to address the problem and um and really get the most out of out of AI.

SPEAKER_00 33:14

You know, Jeff, one of the challenges that we've had is keeping everyone, our customers, even our sales teams and our partner teams up to date on what's going on has been a huge huge challenge. Um and you've been at the forefront of solving that challenge, which is making sure that the story is simple and that story is not something that you do as a marketing activity post the product, but the story is what should instruct the building of the product. That's right. Um any kind of um words of advice to our partners and to our internal folks on like what how how they need to be thinking about narrating um our values to our customers and making sure that we engage our customers and actually doing a co-design effort. Any thoughts on that to close out?

SPEAKER_02 33:57

Yeah, I mean I I think my my main thought would be you know, this this framework that we just talked about, you know, AI Ready Data Centers Future-proof workplace, secure global connectivity, digital resilience, all tied together with AI, um, that message resonates. We've we've um you know, we we hear this constantly from our customers. I think when we can tell the story in that framework, um, and then obviously we need to do the double click. You need to talk about what's underneath that. But when we frame the story that way, um, I think we uh definitely differ become very differentiated in the market and compared to a fluffy story.

SPEAKER_00 34:32

It means that we are able to highlight the fact that this notion of a co-design full stack is something that no one else besides us is able to deliver in the market.

SPEAKER_02 34:40

That's right. And and as as our customers are seeing agentic AI take off more and more and see the challenges associated with that in addition to the benefits, I think they're beginning to realize that you have to have you have to be thinking about all these domains coming together, and that's the power of this the message and the strategy.

SPEAKER_00 34:57

Jeff, thanks for being here, man. Great, thank you.