FinOps in Action

FIA - Gerard Sanchez
===

Gerard Sanchez: [00:00:00] I I think especially for production environments or environments that are super sensitive, there's always gonna be a human in the loop, right?

at this point at least, right, you said we're in the second inning. Might be we're in the bottom of the fifth and these things are just running on their own. Um, but I think human in the loop is o- is ever, at least right now for the foreseeable future, is gonna be very important, uh, for this, this AI age that we're in.

Intro: welcome to FinOps in Action. I'm your host, Taylor Houck. Each week I'll sit down with FinOps experts to explore the toughest challenges between FinOps and engineering. This show is brought to you by 0.5, empowering teams to optimize cloud costs with deep detection and remediation tools that actually drive action.

Taylor Houck: Hello, and welcome to another episode of FinOps in Action. I am really excited about today's guest. He started his career in IT support, fixing printers and troubleshooting desktops, and worked his way up through network engineering and infrastructure build-outs all the way to where he is [00:01:00] today, heading up enterprise platforms and cloud technology at Resolution Life. the past five years, he's built out the company's entire cloud engineering, DevOps, and FinOps practices from scratch. He's been through the qua- crawl, walk, and run of FinOps maturity with the war stories to prove it. Welcome to the show, Gerard Sanchez

Gerard Sanchez: How's it going? Happy to be here. Excited to be here

Taylor Houck: Yeah, Gerard, I'm super excited to be chatting with you, man, and I, I really want to start back in the beginning. mentioned in our prep call that you joined Resolution Life when the company was just six months old, and your architect told you that he literally went into the cloud console and clicked Create Account. Take me back to that moment. What did you walk into?

Gerard Sanchez: You know, it was, it was... I wanna say first off, it was an exciting opportunity 'cause I knew that I'd be building out this infrastructure, this new cloud technology infrastructure from scratch. Um, walked into just minimalist infrastructure build, like very basic, like think, [00:02:00] think user management, things like that.

Firewalls were set up, but the meat and potatoes of the infrastructure, the policies, the securities, the, the servers, the, the functions, everything was, was, was empty. Um, so it was an exciting opportunity to kind of build an environment that you want to from scratch. You know, how many times have we stepped into a company that's been there for 10, 15, 20 years and you go, "Why did they do that?"

Or, "Let's, let's see if we can fix this," and you have to peel back 20 layers of the onion to, uh, to get it to what you would consider good. It was, uh, what was exciting because what we got to build was something that we knew was, was fresh, innovative from the start. Um, and I'm really happy and proud of what we built, uh, to, to, to this point

Taylor Houck: My understanding is that when you joined the company, you guys were in just full on build mode, right? Like leadership's message was essentially, "Don't even worry about [00:03:00] cost right now, just let's build." Can you kind of walk me through the, the thought process and how this thinking evolved over time?

Gerard Sanchez: Yeah. You know, like I said, when you're, when you're building out from scratch and, and you have deadlines and to meet, uh, the last thing you think about is how much it's gonna cost. Well, I mean, not the last thing, but it's kinda lower on the list. I like to say you can do three things. You can do it fast, you can do it good, or you can do it cheap, So you get two of the three. Um, so what we were focusing on was fast and good, right?

Let's build the best environment that we can from the ground up, and we'll worry about the f- the finances of it later. Um, so as we were growing and building and, you know, migrating, um, the, the thought of the, the cost piece of it, the FinOps piece of it was kinda put on the back burner. Um, but then I ran into something called the FinOps Org, and I started really researching into it because I, as overseeing the build-out and looking at the build every day, every week, every month, was seeing our c- cloud costs [00:04:00] like go up and up and up.

I'm like, "Man, there has to be a way that we can optimize this." Everything's consumption-based, you know. Um, this was my first real, like, experience in the cloud and I... But I knew that optimization was gonna be key if we were gonna build a sustainable, uh, infrastructure. So I started looking out there and I came across the FinOps Org.

And it was funny because they were kinda in their infancy too. Um, they-- And it's been amazing watching that organization build out over the years and their first FinOpsX are down in Austin, and now they're out in San Diego. That's, that's an awesome story too. Uh, but really digging into that and understanding like the principles of FinOps and, and how you can, uh, use that to dig into your environment, do cost forecasting, do cost analysis, and, and make, uh, engineering decisions based on those

Taylor Houck: I want to just make it clear to the audience that, Gerard, you come from a hardcore engineering background. This is correct?

Gerard Sanchez: Yes. I come from traditional infrastructure, [00:05:00] cabling system, racking systems, you know, uh, counseling into switching and routing and even servers and, and, you know, with my, my laptop hands-on physical. So transitioning from that all into a cloud environment where everything's, you know, virtual and you can't touch it and feel it was, was a big transition.

But

Taylor Houck: I found through talking to so many people that have gone through this journey of kind of like realizing that FinOps exists and diving deep into it, there was some sort of a, a moment, right, where you recognized, "Hey, this is really important." Maybe it's a surprise bill or a, a cost that spikes, uh, unexpectedly.

Did you have a, a moment like this that kind of put into perspective how important it was to focus on managing the cost of the cloud?

Gerard Sanchez: Oh, yeah. Uh, I remember one day, uh, I got a phone call or a text message or some site of-- some sort of alert notification from my, my cloud account team, like our salesman at the time, "Hey, Jerry, [00:06:00] did you know that you have this serverless compute function that's racking up thousands and thousands of dollars?"

And I was like, "No, but thank you. I'm gonna look into this." Right? So, I know that's not tons of money, right? It's a fraction from what some people spend. But when your full month bill was sixty thousand dollars and you spend ten thousand dollars on a function in one day, like that's a red flag. Um, and we-- That was a really good learning experience for us because no one knew, right?

We had this rogue function running that was just looping through and, and no one was alerted, no one knew it was running. Thank goodness our cloud team had some type of monitoring, and they, they were, you know, we were, uh, we partnered with them enough to feel comf- for them to feel comfortable to reach out to us.

Um, but then you go, "Okay, how can we have prevented this? Like, how could we have known sooner? How can we fix this problem?" And we dug into it, and we saw that there's like budgeting alerts and billing alerts, and we had the ability to [00:07:00] throttle or completely shut down these serverless functions based on certain criteria and build automations around that.

So not only are we getting alerted if something is going rogue, we can actually shut, shut it down and stop it and, and prevent that large consumption bill before it even begins

Taylor Houck: Oh my gosh, I've, I've been through very similar experiences, and it's interesting because especially w- with serverless compute functions, you could have it where it spins out of control and it doesn't only impact the cost for that service, it actually can cascade into other services as well, right? And like what you mentioned too is that you guys didn't even necessarily catch it yourself.

You had the account team that came to you and said, "Hey, did you know that this bill is, is kind of rising like crazy?" And that kind of puts into your mind, "Hey, we need to have systems in place to help us find this when it happens and, and hopefully even avoid it happening in the first place."

Gerard Sanchez: I mean, that's, that's the golden star, right? Let's... One thing to get alerted on it, but stop preventing it or stopping it, stopping it from happening is, is where you want to get to

Taylor Houck: So can you tell us a little [00:08:00] bit about kind of what, uh, has happened or what has, uh, transpired since, you know, this, this sudden cost spike and how you guys are managing the, the optimization and the of your cloud expenses today?

Gerard Sanchez: So I think you mentioned, uh, we've kind of gone from this journey of cloud walk, run, uh, and in the early days it was crawling, putting in alerts, putting in, putting in budgeting, knowing we're tagging. Tagging is so important. Um, if you don't get that right, you're never really gonna know where your spend is coming from, what applications are doing and what teams are causing it.

Um, so really that crawl walk into more of this or crawl into more of this walk where now you're using the data and the, and the framework that you set up to, to make optimization decisions, right? Um, you know, hey, let's, let's look at our, let's look at our ETL jobs and see if we can right size some of the workers.

You know, this thing's only running for [00:09:00] sixty seconds. Well, it really takes thirty seconds to, uh, spin up. Why are, why are we using twenty workers on this to, to run this data, right? Um, a lot of waste there. Then you can go into the, what I would consider the run phase is let's prevent that from even going into our environment to begin with.

Let's shift that left into the architecture and the, the, the, the, the design phase of, of the build out so that we can look at these configs before engineers deploy them into our environment and ensure that they're going in optimized instead of kind of trying to catch it afterwards, right?

and that, and that's kind of where we are today, right? Um, we've shifted left that FinOps practice so that we are as much as possible preventing that waste from going into our environment to begin with. Um, we still continuously optimize. We can still continuously check across the environment to ensure that, that, you know, things [00:10:00] are as optimized as be, as can be.

But that shift left has really made a dramatic difference in, in our overall environment

Taylor Houck: One of the things, that always comes to my mind when people talk about shift left with respect to FinOps is the type of persona or the talent that's needed to really put that into place. And I know that you come from an engineering background. what is it-- is the makeup of your FinOps team, if you have one?

Is it a dedicated team? Do you have shared responsibility amongst engineers? And how technical are the folks that you have thinking about, uh, how cost is related to how you run your cloud environment?

Gerard Sanchez: So when I was thinking about how we should position FinOps within, uh, within Resolution Life, I really came to the determination that we need to get FinOps into the hands of the people that can make a difference, right? The people that understand these configurations, that understand how these applications are, are deployed and how [00:11:00] they're, how they execute on whatever processes that they're performing.

we do have a dedicated FinOps team. It's a small FinOps team, but it is dedicated. But that FinOps team sits within our CCOE and our engineering vertical. So they sit closest to the engineers that help manage and build out the infrastructure of our environment and can understand and talk to the application teams and, and can talk technical with them so that not only do we understand how the, the, the environment is working, the application is working, but the application team can understand when we're making recommendations to them in a technical, in technical terms that they can go and actually execute on or we can execute for them on their behalf.

Taylor Houck: It's so

Gerard Sanchez: So-

Taylor Houck: to have people that have the ability to really understand what the impact is of the recommendations that they're making or the im- or really understand the real system design of the applications that they're working with. Because, I mean, I, I've seen in my experience, and maybe you've seen the same, that oftentimes you might have, you know, obvious [00:12:00] cloud-native services that people are using, but that's wildly over-engineered for a specific use case

Gerard Sanchez: Uh, yeah. Um, you know, this perfect example here, we were, we were throwing... I don't know if anybody gets car references, but when we were looking at our user shares, how we were gonna build user shares out in this cloud environment, at first we were like, it was actually my infras- my engineering, my, my team, not even an application team that said, "Let's use this certain type of, of high performance storage, uh, on...

to, to build this out," right? At, on the cusp of it or at first look, "Hey, this is great. Super fast, super performative. Users will never have to worry about the performance of this." And then you get the first bill, right? And even though you do all the calculations beforehand, it really is kind of like a little bit of a sticker shock.

So then you go, "You know what? Do users really need to be running on like, uh, 10,000 IOPS SSDs?" Probably not, [00:13:00] right? Like, what is that real user experience when they're opening up an Excel file? Right? So you look at other... You look and you go, and you, you, you work with your, your cloud partners, your, your cloud team, your cloud, you know, uh, cloud partner team, and you research and you say, "There's this, maybe this significantly less expensive," which still, uh, can exe- you can still execute the use case on, uh, where maybe you use bucket storage as a backend.

A little less performative with throughput with the front end system on it, but you're still providing m- far better experience than, or far, far more performance than an, an average user needs, and it even scales, right? So it's, it's important, right, to look at the workload that you're actually solving for when you're looking at the cloud services that you're using to execute on that, right?

That, that specific use case was done before we, we pushed the FinOps practice less, left, right? We were still kind of in our, our walk stage. [00:14:00] Nowadays, we would've caught that earlier, right, in the architecture and design phase and said, "Maybe this, this, this, uh, solution is a little overkill. We're throwing a B- Bugatti at it.

Maybe we could just throw a Toyota Corolla, you know?"

And it would be just as fine

Taylor Houck: sitting on-- I, I live in, in Connecticut and we have I-95. It runs right through my town. And pretty much every time I get on this road, we're sitting going 35 miles an hour on the highway, right?

Gerard Sanchez: Exactly

Taylor Houck: could be in a Ferrari or a, a, you know, 1997 Toyota Camry and I'm, I'm gonna get where I'm going at the exact same time.

Uh, now if you're on the racetrack, you, you want that, that nice car. But on 95 in, in Connecticut, no, you don't

Gerard Sanchez: No,

Taylor Houck: it

Gerard Sanchez: knowing your use case, right? Knowing your use case and solving for it

Taylor Houck: so let's really dive into this because I, I-- You've talked about the wake-up calls, right? And you've talked about building some guardrails in place and getting together a team and getting the right people involved. Can you talk about that shift from kind of playing defense and being more reactive to more of that [00:15:00] offensive and proactive approach to getting FinOps involved in the design phase?

Gerard Sanchez: Yeah. Like, like I mentioned earlier, it's, it's all about shifting it left into, you know, the dis- you hear shifting it left and, and some people think shift left and you're putting it in the application team's hands, right? Giving them the visibility into their applications and the data and the cost data so that they can make decisions and see those decisions in almost real time of con- their consumption costs.

I look at it as a two, two-pronged approach, right? Not only do you push that and give that visibility to the application team so they understand real time or near real time what, what they're deploying into the environment, the cost of those decisions and...

But not only that, you, you push your FinOps practice left, and again, because they're engineers, because they understand, uh, the, the designs and the, the configurations of, of the application, they can help the application team [00:16:00] make cost-efficient, performative decisions, right, before it even goes into the environment.

So you can say cost prevention and also visibility to the application teams for what their application is actually doing

Taylor Houck: Really interesting. And I, I wanna touch on two, two points here that I hear a lot, especially where we are right now in 2026 as it pertains to FinOps. we've been talking about for a while, and that's automation, right? And using automation to help improve your FinOps practices. And the second is AI.

Gerard Sanchez: No to that

Taylor Houck: of use cases around how AI can not only, you know, expand the scope of your FinOps practice because you're seeing a lot more spend in different areas and new pricing models and new ways of thinking about how do we manage and, and quantify the value of the investment that we're all making in AI, but also how are we using AI to make sure that we are truly running a modern FinOps practice in 2026? When I, I bring this up, what, what comes to your mind? How are you thinking about these problems?

Gerard Sanchez: [00:17:00] So first off, you wanna make sure that you have the visibility when you're, you're, you're building out this infrastructure and framework for teams to deploy their, their agentic and generative AI solutions to. You wanna make sure again, that those teams have the, the ability to see what those, those costs are, right?

Um, you know, auditability, see what's actually running, what these, these agents and these generative agents and agentic agents are actually running, right? And then y- when it comes to using it, right, making sure that the use cases that you have have access to the data that it needs to make the right decisions and, uh, execute as you, as you design it to, uh, design it for.

So personally for me, within the FinOps space, it's, it's been like a, a revolution, right? the tools that are out there that you can now use to, to d- perform analysis, deep dive analysis into your environment, into your architecture, into your engineering decisions and configurations, [00:18:00] it's, it's not only is it a sh- huge time saver, things that used to take days and days of looking at numbers and looking at configurations, you can run in minutes, right?

So you can make almost these, these quicker, faster decisions, smarter decisions, as long as you, one, prompt correctly, right? That's a big one. And two, give these tools the access to the data that it actually needs to make these informative decisions

Taylor Houck: it's truly transformative, and I'm seeing so many people across the FinOps ecosystem talking about the unique ways in which they're using AI to improve the velocity of their FinOps program in order to get people answers faster and quicker. I'm curious to get a little bit tactical with you here, because I think it's, you know, a lot of people are talking about, oh, AI, it's here, it's valuable, it's huge.

But to get very specific, what's the architecture of your FinOps AI use case? Like, how are you actually using these tools and plugging it in to get the, the data that you need?

Gerard Sanchez: Yeah. So I think first off, make [00:19:00] sure you have to-- First off, you have to ensure that you have the right permissions, right? Not over- overly permissio- permissive, right? The, the roles that you build, you don't want them to have execute really on anything. Let's be safe here. Let's maybe only read-only, right?

You don't want your tools accidentally, we've heard about it in the news, right? Accidentally making decisions for you and executing on them. Next thing you know, have your production environments down, right? Because you were running a FinOps analysis or an efficiency engineering analysis on some of your use cases.

but then, like I said, you ensure that you're... Like for me specifically, I use an agentic IDE, right? And I use, uh, MCP servers that-- to then attach my agentic IDE to the data sources that it actually needs, right? So think I have this super smart agent that now has access to my engineering documents on a read-only basis.

It has access to my user stories and tasks that my [00:20:00] engineers execute on and, and, and des- or, you know, build, create when they're deploying stuff into our environment. Then it has access into my actual cloud environment to v- to view all the cost data and the, the logs and the application logs and the, and the security logs and execution logs, and you can put that all together, right?

So a perfect use case that I use agentic AI for recently or the, the in a FinOps is to analyze our data lake and optimize it from a data storage and data backup, landscape. So think I have this tool that can go out and look at the architecture of my data lake, and I have the, the-- this tool that can go out and look at all the stories and tasks that, that were used, that were created when building out this data lake and understand how the flows work, how the configuration works, how data is processed through this [00:21:00] lake, and make informative and intelligent decisions on data that doesn't need to be stored long term, data that doesn't need to be backed up, right?

Um, and, and give you, and give you reasonings why, right? In a day where that type of deep dive analysis would take weeks if you were to do it by hand, right? And say, "Hey, this bucket doesn't really need to even be backed up. This bucket, you don't even need to store data for more than thirty days because it can rebe-- it can be reprocessed easily or, you know, a new batch comes in and you don't need it."

So having this agent that can use the data that you give it access to. One, the da- that's why I keep saying the data that you give it access to is so important, um, and then create and reason, give you reasoning behind why you're making these, uh, configuration and engineering decisions

Taylor Houck: Yeah, and you can give it access to even, hey, if you have tags in place, this is what the tags mean, right? So it can then put in some context from your specific, um, [00:22:00] business context. But I, I wanna ask an important point here because I think it's really, really important when we're launching, you know, these parallel agents to do analysis on our behalf, and it would've taken weeks and now it's taking, you know, minutes or hours. How do you trust and validate? Because if you're gonna go and, and validate every single thing that the agent's going to, to tell you, that's gonna take up so much time that it's, it's like, what's the, the benefit in the, the first place, right? So how do you think about managing, especially, you know, from a, let's say a management perspective, right?

I mean, you are a leader within your organization. It's not just the analysis that you're doing. You have teams and people that are doing analysis using AI, and do you just trust it at face value? How do you think about this problem?

Gerard Sanchez: Uh, I think this all goes back to where these type of decisions and vertical, you know, this analysis is being performed. It's being performed within the engineering, um, vertical, right? So it's not somebody off in finance or, uh, you know, whatever, looking at this and just [00:23:00] throwing things over the wall to the engineering team to make decisions.

The, the people that are performing these analysis and executing on these analysis are people that understand, right? And they can look at a solution or look at it and be like, "You know what? I don't-- I, I helped build this out," or, "I have the documents myself and I don't think this is right." and ensuring that you feedback, feed that back into the, to the agentic AI or the, the IDE that you're using or the generative AI IDE that you're using and, and use it to, to create smarter, more intelligent answers.

Um, I can't stress enough how, at least from what we've it needs to sit within engineering, right? Tho-those decisions need to be understood, um, and analyzed before being executed on

Taylor Houck: How are you thinking about managing the cost of AI? Because, I mean, you're mentioning, "Hey, we are deploying and giving our engineers access to these tools. They're doing all this analysis. We trust them to make sure that it's valid and accurate." I'm sure they're munching a lot of tokens.

Gerard Sanchez: Mm-hmm.

Taylor Houck: you thinking about [00:24:00] managing the cost of AI?

Gerard Sanchez: Uh, first thing is the visibility. It all goes back to visibility, right? Um, ensuring that things are tagged appropriately so you can, uh, track them back to the use cases that they're performing. And let's be realistic, not all AI use cases are beneficial, right? You might see that it costs-- Like I think one thing that people are starting to see now is the cost of building some of these solutions far outweighs, or the cost of running these solutions far outweighs the use cases they're solving for, right?

So visibility, uh, tracking, and then ensuring that the way you're using AI, um, is optimized as well, right? And that goes back to the visibility of looking at prompting, how you can optimize prompting. Um, and then just following your, your FinOps, uh, practice, right? Visibility, giving the data, making decisions based on the data that, that you're, you're collecting

Taylor Houck: Where do you see this going? Because I think that almost everyone would agree [00:25:00] that we are in the first inning, maybe we're in the top of the second when it comes to this AI wave. How transformative do you think this is gonna be? You know, and you, you can almost zoom out from just the pure FinOps lens, right?

I know you manage much more than just FinOps. do you see AI transforming the way that companies are run in the future? what is then the impact, getting back to FinOps, cost of AI, it's almost, in my mind, like a new HR in a way, because really we're gonna be augmenting labor and, and using technology to own tasks and outcomes. I, I, I'm just interested from a, an executive perspective, where is your head at when it comes to the future of business?

Gerard Sanchez: Oof. That's a tough one, right? Because who knows how far we can-- how far this is gonna go. Um, I think from my perspective within the technology vertical, I think we're going to get to a point where a lot of these agents will be [00:26:00] able to make informed and, intelligent decisions and automate the execution of that.

Um, I don't think we'll ever really get... I mean, maybe eventually if the models and the intelligence gets good enough, we, we will get to like a fully automated. But I think especially for production environments or environments that are super sensitive, there's always gonna be a human in the loop, right?

because you just, at this point at least, right, you said we're in the second inning. Might be we're in the bottom of the fifth and these things are just running on their own. Um, but I think human in the loop is o- is ever, at least right now for the foreseeable future, is gonna be very important, uh, for this, this AI age that we're in.

these AI use cases are gonna be far outside of my technology realm, right? These automations are gonna be, you know, used in the business across, across the business, right? Think actuarial use cases, investment use cases. You already see some of these things where have [00:27:00] AI that are making investments decisions for people, and they're, they're doing phenomenal jobs, right?

Um, again, human in the loop is gonna be imperative, but I think it's gonna transform a lot more than just the technology space which we're living in. It's gonna transform the entire business landscape. And it'll be interesting to see how far it can go before we decide that humans aren't needed anymore to perform those functions, right?

Or to, to validate and trust and, um, have that human in the loop piece. I don't think, at least, I don't know. I don't know if it'll get there in my lifetime, probably, definitely, but not in any time in the foreseeable future

Taylor Houck: Yeah. I mean, I think that it, it comes down to this, right? And this is again, an opinion. Um, but you talk about human in the loop, of course, right? We need to steer these agents, but what it's going to enable us to do is act at a much greater scale, right? So it's almost like one of the ways that I've been thinking about it mentally, and I'd be interested to hear, you [00:28:00] know, your, your feedback or your thoughts on it, but it's like you take every individual contributor world, right?

And we can start with engineers because they're the most, you know, they, they have the right mindset to embrace this, but even more, let's say, future-looking or, or, um, openness to learn that people from other divisions or departments in a company are embracing this as well. Every individual contributor is now almost akin to a manager because they can have a team of agents that are doing work on their behalf.

Now, they need to instruct these agents. They need to tell the agents what to do. But if they can construct a system that is, you know, working properly, then they can have the power of like a manager, right? And then now you take the manager, the manager has a team of people, and the people have their teams of agents. So now the manager is more akin to a director, right? So to speak. And

Gerard Sanchez: Right

Taylor Houck: kind of gets leveled up in a way. Right now, if you push against this trend and say, "No, I'm focused on myself," hey, you can go still, you know, provide some value in the marketplace for [00:29:00] sure. But you're, you don't have this, this leverage on yourself, right?

Um, that, that's the way that, I don't know, my, my head is at it on, you know, in May of 2026.

Gerard Sanchez: That's, that's a really interesting use case, how you can, you know, use agents and, and keep that human in the loop at a much bigger scale, um, and really just drive optimization and performance across almost the entire org, right? If you can build the use cases. Yeah, that's, uh, you know, let's... It's interesting.

It's fun, right? This, uh... We're living in like... I take this... Think of the people back in the day that were just figuring out what the internet is and, and really building that infrastructure out and building out, you know, building out these servers and going back and s- seeing the people the first time they sent a file across the internet, right?

I feel like we're-- It's exciting because we're living in that kind of that age, this, this new age of technology with this AI, and it's, it's been a really fun experience to learn about it, to start implementing, to, you know, hearing about it. You start hearing about this AI stuff, and you're like, "Hmm, [00:30:00] that's interesting."

And then you start seeing these use cases that start coming out, or you start trying it out. Hey, I remember when ChatGPT came out, and I was like, "You know what? I'm gonna go and check this out." Right? And now going from that to using it in your everyday life, it's, it's been really exciting and fun to see how it's progressed and how it's really, uh, expanded.

The use cases have really expanded and, and caused what I would consider significant impact in the industry

Taylor Houck: Oh, absolutely. I mean, without a doubt it's causing a major impact. And just as we're getting close to, to putting a bow on this episode, Gerard, I'm really interested to hear kind of where your head is at in terms of what's next for you at Resolution Life as it pertains to AI. Are you looking at building internal models?

Are you looking at, you know, going in-- And you don't need to get into any company specifics that gives anything away, so please feel free to deflect this question if you're not Um, but what, what's next for you guys in embracing AI and, and more specifically within [00:31:00] the, the FinOps practice?

Gerard Sanchez: Yeah. I mean, it all goes back to building out the correct framework for the u- you know, for the business to deploy these generative and, uh, agentic AI solutions too. So you have to make sure that you have the right permissioning for the agents. You have to make sure you have the right visibility for the agents.

You have to make sure that you... Now is the time, right? When your company's getting into it and you don't have a f- You know, think when I talked back to you, walk in 20-year company, and you have things, legacy things that have been there. But now it's like, if your company's just starting to get into AI, put the time and effort into building the framework and the foundation as strong as possible and doing it correctly so you have the visibility, you have the tagging, you have the auditability, you have the ability, you throw in your, you're ensuring that the human in the loop is there.

guardrails , Right? You don't want these agents running amok, right? Put in costing guardrails or guardrails, this goes back to permissioning, but guardrails to ensure that they can't get into data that they're not supposed to [00:32:00] or perform actions that they're not supposed to. and I really think that's where we're focusing on right now, is just ensuring that we can build the best foundation.

And we do have some use cases, don't get me wrong. full speed ahead on the building use case, building out models and, agentic, POCs and use cases. But while those are running and they're running in their respective sandboxes, we're thousand miles an hour ensuring that when they're ready to go, we have a foundation that is strong and scalable so that the entire company can leverage the AI tooling.

Taylor Houck: Amazing

Gerard Sanchez: Making co-

Taylor Houck: man. It's gonna be, it's gonna be really fun, and I'd be very interested in following up and maybe having another one of these conversations in months or a year's time and hearing about how transformative it's all been because things are just moving fast. But before we put a bow on this conversation, Gerard, I, I wanna really give our listeners, especially those that are earlier in their careers, opportunity to learn a [00:33:00] bit from you.

Because as I mentioned in the intro, you have really done an impressive job building your career into what it is today and have really kind of jumped through all of the ranks into a high-level leadership position. If you were to go back and, and give advice to either a, a, a former version of yourself or someone who is, let's say, in the, the first couple years, let's say first five years of their career today, what advice would you give to them?

Gerard Sanchez: I know this is kinda like the generic answer, it's, but be curious. Always be curious about what's going on around you. Um, always ensure that you're trying to grow, and even outside the technical landscape, right? Ma- but ensure you're always being curious and trying to grow outside the technical landscape, right?

For example, explore solutions, right? Learn more about what's in the market. Don't just be happy with what you have, right? It doesn't mean that you're gonna shift away from the solutions that you have in place, but learning [00:34:00] more about other solutions in the market or learning more about the technologies that drive those solutions is really gonna help you understand and, and as you move up the ranks, understand how those, those, those things work in the larger organization.

Um, continuous education. Never stop growing, right? Um, I, at my age and where I'm at, I still am pursuing certifications, right? Uh, I wanna always continue to make sure that I am the best person that I could be in the position that I'm at, and I never stop putting... Keep your foot on that gas. Keep that foot on the pedal, going back to all the car references we've done today, right?

Um, and then one of the biggest things that I'm gonna say is find the right mentors, right? Find the people in the organization that, that wanna help you and wanna help you grow into whatever you wanna become, right? S- some people don't wanna go down the path of management and, and, and [00:35:00] executive or leadership and executive leadership.

Uh, some people wanna stay an engineer, right? So find those people that can help you pursue the path that you wanna pursue, right? Um, I can personally say that I don't think I would be where I'm at today if it wasn't for the mentors that, that I've had along the way, and I'm eternally grateful for those.

Um, but find people that are like-minded and that wanna see you succeed, and ensure that you ask the right questions and listen and take that into account and try to build yourself up, using those, that, that experience that they have.

Taylor Houck: That's amazing advice, Gerard, and, and I think that everyone listening has something to learn from that. As you mentioned, it's not only for people that are earlier in their careers. Everyone should be thinking about, you know, things from a growth mindset, and it's especially important as we were just wrapping up a whole conversation about AI. There's no one out there with 10 years experience of managing these agentic systems within enterprises. There's

Gerard Sanchez: Nope

Taylor Houck: out there. So you don't have people that [00:36:00] have this huge, you know, leg up on you, even if you are earlier in your career, even if you don't feel like you have all the experience in the world to go and drive an impact. Hey, there really aren't that many people that do. And with the pace in which this thing is coming, gonna need folks that are at the forefront. So, uh, I really appreciate that advice. I think it was excellent advice to anyone that's listening. And just as we, we put the final wrap on this, Gerard, for anyone that's listening to this episode, if they're interested in, you know, learning more or getting in touch, the best place to find you?

Gerard Sanchez: reach out to me on LinkedIn. It's, uh, Gerard Sanchez. Um, I think I'm the first person that pops up. I don't know how many Gerard Sanchezes are out, are out there. But feel free, if you want to reach out and, and talk to me about my experience or just chat about FinOps or technology or anything, uh, reach out and I'm, I'm always ears.

And like I said, talking to more people and learning, right? Uh, getting different, uh, viewpoints is, is, uh, a great way to, to grow

Taylor Houck: Awesome. Gerard, this [00:37:00] has been incredible. Thank you so much for coming on the show. I really appreciate it

Gerard Sanchez: Thank you. I really enjoyed it and, you know, hopefully I'll get to talk to you again in the future

Taylor Houck: Absolutely. And thank you to our audience. If you learned something today or got something out of this conversation, which I'm sure you did, please share it with someone who needs to hear it. This has been another amazing episode of FinOps in Action, and we'll see you next time

Outro: That wraps up another episode of Fit Ops in Action. Thank you for joining. For show notes and more, please visit fit ops in action.com. This show is brought to you by 0.5, empowering teams to optimize cloud costs with deep detection remediation tools that actually drive action.