Episode Transcript
[00:00:00] Speaker A: Welcome everyone to a new episode of the FinOps weekly podcast. And today we are going to talk about incentives and you know, how organizations can be better using finops with those tricks and how to involve everyone. And for doing that we have Tatum Danix, the senior product manager from Kion. Tatum, how are you doing?
[00:00:21] Speaker B: Great. Thankful to be here. Looking forward to the conversation.
[00:00:25] Speaker A: Yeah, pleasure to have you here today.
You know, it's, it's always great to, to talk about the incentives and the different things that we need in, in finops. You know, from, from the engineering side, definitely incentives, it's someone. Something super important in general to, to phenom. So what do you think are the most common like incentives that sabotage finos program because they are bad before they like even start.
[00:00:53] Speaker B: Yeah, so I think, I think the most common one is especially at the leadership level, folks start to prioritize just dollar saved, right? Everything in FinOps becomes about dollars and cents saved. And what that does is you start to create an incentive and a culture around ignoring the rest of finops, right? The rest of the way that you can create value. You know, you start to ignore, you know, unit economics and it's just about how much money are we saving? And it's not rooted in is this the right amount of money we should be saving? Is there more on the table and it just creates a bad culture around what finops is ultimately supposed to be.
[00:01:29] Speaker A: Yeah, definitely. And I think that we always start with the idea in FinOps, I think with the cloud cost optimization. So like reduce cost and that's where you are introduced that.
But then you are always evolving once you start, especially from the business side and try to evolve to a more business value. So cost is not bad whenever they are related to business value. So if you are increasing your cost but your margin is increasing, then let's bring it, let's bring it up.
[00:02:00] Speaker B: Right, exactly. You know, a simple, you know, in a prior life, before being on the vendor side, I was a practitioner at a healthcare company for many years.
And you know, one of the simple things we did, you know, part of our business is based on transactions, medical claims, things like that.
That typically is a good unit to measure our cloud cost in. Right. Like cost per transaction or even just something as simple as cost per revenue started to really change how our leadership looked at our cloud spend. So simple things like that are a good way to get started. But there's definitely still a lot of practices out there that are incentivizing the wrong thing.
[00:02:38] Speaker A: Yeah, yeah, definitely. And you know, I Want to dive deeper, more on that because you may have like, you know, that's coming from Kieran. You may have like knowledge from, from the organizations and the clients you work with. So what do you think are some things that organizations like clients that you may have, are doing today that are like certain games that, you know, setups for long term favor and how you are able to fix them for them?
[00:03:04] Speaker B: Yeah. So what folks are doing right now, and this has been typical for many years, is you get in this mode where everything is reactionary. You.
There's no, there's not for lack of dashboarding and insights across any tool or any even homegrown solution you might be using or even native experiences. Right. Folks can pretty much get insights into where they might have waste or things that they can, you know, start to clean up. And when you're constantly spending all of your energy thinking about how to take action on those and you're not doing any thought on, well, how do we prevent this from happening again or how do we at least shorten the time before we're actually able to take action?
And so what folks are doing is they're in that fight or flight reactionary cleanup mode all the time. And if they would put just a little bit of that effort into thinking more strategically and proactively.
[00:03:57] Speaker A: Right.
[00:03:57] Speaker B: They would pay dividends over the long run. And, and so that's what I try to have conversations with folks on now is around, you know, what policy and governance can do. Because what I'm trying to do, both as you know, on the vendor side, but as someone who just loves this industry and grew up professionally in this industry, I want to help people be successful and try to make them realize, going back to the incentive conversation, that a dashboard showing dollars going down isn't necessarily showing the true value that you could be bringing to the organization.
[00:04:27] Speaker A: Yeah, yeah, definitely. We have like examples on other podcasts that talking about the context that it means a lot, the context in the conversation rather than just numbers and what, what does it mean for you? Like having a number on itself is like, it can be great, it can be worse depending on if it's business, like it's benefits or is it cost. So you know, you need to give context on that.
[00:04:50] Speaker B: Well, and a great example of that. So this is a true story. The first year of FinOps X, this is all before coming to Kion.
I was having lunch at one of those tables where you're kind of just figuring out where to sit. And I sat down next to and we started to Strike up a conversation. This is still like finops is still, it's around, but it's still in the early days.
And he starts sharing with me how his CIO is so in love with this savings dashboard that he built himself that he had actually started to purposely not clean up things.
Because what had happened was this is typical, but that first six months, kind of the honeymoon of FinOps, where you have all this opportunity, you can buy savings plans, reservations, all the, you know, the charts are going down, down, down. And so leadership gets excited, right. Look at all the money that this team's saving us. And the well had kind of dried up, if you will. And so what he started to do is purposely leave things that he was finding as opportunities to help showcase savings, because in his mind, right, that's his job security.
And so that is the, the type of thing that I hope we're doing less of now that we've advanced, you know, four years since that conversation. But it still definitely exists where folks are struggling to show how else they're bringing value besides something like a dashboard of savings.
[00:06:12] Speaker A: Yeah, definitely. And you know that you need to put the incentives in the right context and in the right objective. Because if you are incentivizing about the objective, people are going to trick, as you mentioned, the system and try to, you know, figure out that is good for themselves. Right?
[00:06:29] Speaker B: Yep, exactly.
[00:06:30] Speaker A: And, you know, can you, can you set an example where, where like fixing the incentive or changing the way that an incentive was planned, fix the, the cost or the problem without like knowing anything else apart from. From that?
[00:06:43] Speaker B: Yeah, so I can give an example of that. So there used to be a VP of engineering who, you know, many years ago, I would say was pretty forward thinking with finops. Kind of understood it before it was even kind of popularized. This is even before the foundation really had taken off.
But his team specifically was really struggling with just your common use cases, Right. They weren't tagging anything, they weren't shutting off anything on the weekends. You know, the, the common, simple examples. And what actually happened on that team is one of the more junior engineers started to just take some initiative, started to writing, write some scripts to clean up tags, started sending Friday emails with kind of an overview of spend, and started to do kind of some grassroots finops things, his own.
And so that VP ended up promoting that junior engineer to basically highlight the efforts that he was bringing to the team. And what that did is it started to establish kind of a culture.
[00:07:40] Speaker A: Right?
[00:07:41] Speaker B: This is an important way. This is a part of your job. This is part of the expectation of what being an engineer looks like. It's not just coding and innovation, it's also proper hygiene and being a part of this larger finops process. And so that's a way where you kind of change the culture somewhat overnight when you start to show people that there's a career path here and advancement here by just caring a little bit. Right. Maybe it's 10% of your job or 5% of your job thinking about the things that you're doing as an engineer and how that contributes to spend and value.
[00:08:12] Speaker A: Yeah, definitely. I think it's a great case of a great example of that. You just need to sometimes fix the way the problem is design or plan and then you just have to switch and you don't have to change the whole framework or you need to re architect everything or some of that. It's just people need to work the process. Let's say the process that you need to put your FinOps team into is like it needs to be well targeted so that the goal is clear so that everything works right. Like any other team in the business.
[00:08:45] Speaker B: Yeah, exactly.
[00:08:47] Speaker A: And you know, moving on to more of the incentives, I wanted to also talk about on the AI side. So you know, there is a lot of conversations of course, in the AI for finance. Finance for AI. So like from your perspective, from the KION side and maybe for your personal one, what do you think is the most impactful use case for finops that AI has for finops that is working really in production at scale right now?
[00:09:16] Speaker B: Yeah, I've got a lot to say here, so you chime in anytime, have a comment on something here, but I think there is a lot of noise and I actually think the better part of this year will continue to be probably more noise. I'm hopeful at the end of this year we start to.
Folks can kind of see the bigger picture of what's going to stick in terms of an AI for FinOps use case.
But I'll give you three that I think exist already today and are going to continue to stick around.
The first is actually around usability. So if you think about a less mature finops team, or maybe it's just an adjacent Persona like your accounting team or your procurement team or someone whose day job is not finops.
What we're already seeing AI do is make it easier for those folks to contribute to FinOps or get the answers that they need. And so what I mean by that is, you know, let's say you've got a tool at your organization that you wish, you know, the accounting or finance team would take more self service actions but they just either never had the time to learn how to use the tool or finops and cloud spend is this big scary thing and they don't know how to get started or how to like be a part of of the process. Right. What AI is allowing is whether it's through simple canned prompts that folks can click on or it's just using simple language like they're used to in ChatGPT or whatever they use in their personal life. It's giving them a place to just ask simple questions and learn how to get the data or the recommendations for next steps or whatever it is that they should be asking in a much easier, less scary way. And I think what I'm seeing is really cool there is it's actually creating more of a finops culture because people now know how to go get information or ask questions on what does this data mean to me, what should I be thinking about and all of that. So that's one use case that I've seen really take off and I'm excited about going forward.
[00:11:11] Speaker A: Yeah, definitely. And just to add on that is like I'm very used to and probably you as well like to see like on Freenaux there is a lot of data. So you have like a huge amount of data and huge amount of, you know, the usability of data like in a database is difficult. Like understanding how you can use that data to get what you want for like to get the insight is very difficult. So that's definitely like it improves a lot the usability of the data. So the insights that you can get is so quick to get insights instead of you know, you doing your calculations and thinking about how to build the dashboard and all of that. I don't know if you agree on that.
[00:11:50] Speaker B: Yeah I do and you know a prospect that I'm working with right now.
One of the things that they are excited about in terms of using our tool and you know we're, I like to think that we're forward thinking in this way but we're not the only tool that does this either. But our in app agentic experience we have these canned prompts and so what they're excited about is all they need to do is teach some of those adjacent Personas on which prompt to come in and click and in one click, right they kick off an entire experience of analyzing spend taking all of that data that they don't have to worry about how to dimension or how to Filter or whatever it may be and it kind of gives them an easy to read format to go forward in their job. And so I think we're going to see a lot of folks start to use that as a way, a lever to bring more folks in to the finops processes.
[00:12:42] Speaker A: Yeah, definitely. And okay, so what's your second one on the AI side?
[00:12:47] Speaker B: Yeah, the second one is probably the most obvious to folks because this appears or this is a common thing in any industry, but it's efficiency. But specifically, I think where we're seeing and going to continue to see efficiency with these kind of agentic experiences is around simplifying those little more administrative tasks that you do in finops that you don't need to waste an hour or 90 minutes on a week or in your month anymore.
What we're seeing is that I'll give you an example of we have a customer who has about 100 different cost centers and their process is every month they come in, they pull some reports and they go to a dashboard that says, hey, budget, forecast variance, what's the difference? And then they go through and they actually update all of those budgets now with the new run rate and new forecasts and communicate that back to finance.
Well, now what they're able to do is a process that takes them about 90 minutes. They ask a simple prompt around, can you give me my cost center's budget to forecast variance comes out with a report. They can review that report very quickly and then actually ask the agent to go in and update all of those budgets to meet where the new kind of forecast run rate is. So now it's active in the tool so anyone can log in and see those updates. And they did it all in under five minutes. And so now they've gained over an hour of their day and their time back to go do something else with more value. And so it's finding those little things that right now organizations know they're probably spending a little bit more time than they should be on. And if you find a couple of those very quickly, you've now given yourself or your teams hours back in their day or in their week and that adds up over time. And so that's one that it looks a little different in every organization, but actually relying on kind of an agentic administrative action type of approach is here to stay for sure.
[00:14:43] Speaker A: Yeah, definitely. I think it can be applied to almost any industry with, you know, adapting the, the functionality to the specific use case and in freedoms I think for sure, like automation. And you know, it's one of the things that improves the scalability, like forever. And with companies I don't know that have thousands of employees each like hours saved each hours, if it's in a long department on a big department, it can save like a lot of money. Like a lot of money, meaning human resource time, which is like one of the most costly spends that we have because we are always talking about the cloud. But the human resources spend is definitely one of the highest drivers.
[00:15:29] Speaker B: Yeah, absolutely.
[00:15:32] Speaker A: So do you have another use case that you want to tell us about the AI thing?
[00:15:37] Speaker B: Yeah, and this one is a little bit more advanced or could be in terms of maturity, but I think over the long term it's where I'm most excited about what AI is going to bring. And that's actually around improving your governance. And so this can be governance of AI spend, but it's not specifically a finops for AI conversation. This is governance in general for your finops practice. And, and what I mean by that is typically the idea of creating a policy and then putting automation or guardrails to make that policy go work on your behalf can be a technical challenge. And if you're a FinOps person or a FinOps team that doesn't necessarily have that technical expertise or that background, it can be a little challenging to go out and build or do that on your own. And so what we're seeing, and we're going to continue as a platform to invest here and is this ability for AI to write usable, correct policy and then know what to do with that to create automation. So for example, one of the things that we're doing and we're seeing customers actually use already is we've actually written a sub tool or kind of a sub agent experience where it knows how to write Cloud Custodian, which is a policy as code or policy engine that folks, it's pretty common across many companies. So it knows how to write Cloud Custodian. Now through a simple prompt where I say, hey, I've got a bunch of unattached EBS volumes, can you create a policy for every time one's been unattached for more than seven days? You either alert this team or if it's in a sandbox environment, you delete it. Whatever that policy is, whatever makes sense for your organization, you ask that question. And now it can create an entire policy that's ready to be implemented and used without any sort of technical expertise. And so it speeds up that time of actually getting real governance and real automation. And I think that is going to be very powerful. As teams mature and are ready to kind of take that next step.
[00:17:42] Speaker A: Yeah, no, I totally agree. I think governance is one of the, as mentioned, along with automation, they are fairly related. And you can scale a lot your finops through governance because it affects everyone. You can automatically apply things to all the Org, which is like, you know, the most scalable thing to do and you can do, you know, a lot of, a lot of impact.
Okay, so we've talked about three use cases where the AI is pretty useful.
Do you think there are cases where the AI can generate more problems than solutions in, in FinOps?
[00:18:21] Speaker B: Yeah, I think if you're using AI as just another way of getting insights, you're wasting your time and money.
I think the industry has already taken big strides in giving folks and practitioners enough opportunities, enough dashboards, enough insights. Right. And I don't think that if that is your only use case for kind of an AI for FinOps type of feature or tool set, whatever it may be, you're kind of missing the potential, right? This, the actual agentic or an agent doing actions on your behalf, that is the use case and that is the long term value you're going to get from features or tools like this. And so it can be helpful, like I talked about earlier on, just getting started, especially for other folks who aren't in FinOps or you're brand new to FinOps. That simple prompt of tell me about my spend is important in those cases. But if that is your only use case, I think you're eventually going to find that the money that your organization is spending on those, on those prompts and those outputs are not justifying the value you're getting.
[00:19:29] Speaker A: Yeah, definitely. No, I agree. I think, you know, AI like it can be really helpful, but you need to use that, right? You need to provide always as mentioned, the right context and the right inputs.
And with AI, the amount of context and the amount of confusion that it can get is super, super high. Because there is a lot of things like you can include your whole company code base and that's it. And that would be a lot of context and probably the answer wouldn't be really useful because there is a lot of code in there. Right.
It's difficult.
[00:20:02] Speaker B: Yeah, exactly.
[00:20:04] Speaker A: So talking about fails, what do you think are the most common or mistakes that happens when companies like when they, before they arrive to you or to the tool, like what they are doing traditionally in the cost management side and what do you think is the root causes that you are normally encountering on those issues?
[00:20:25] Speaker B: Yeah, so there's two One is kind of high level and then one specific.
I think one of the issues or mistakes that I see organizations make with kind of traditional cost management is one they, they actually overlook the impact of culture. And I understand culture as a big loaded word that can mean a lot of things.
But what folks don't kind of grasp is that finops is ultimately a team game. And you might have a team or a person that's in charge of driving or facilitating or, you know, essentially responsible. But if you don't have a culture around your engineering teams, your leadership, whatever it may be, and they don't all see this as part of their job, I think you're set up for failure kind of from the beginning.
The second piece is a little bit more specific and I touched on this earlier, but the focus on reactionary versus proactive. And so this idea of cost management is go find an opportunity that exists and remove that opportunity, right? Right size or delete that piece of waste or tag that instance.
You have to start to shift from the very beginning and start basically incentivizing to go back to that word on processes that are repeatable, right? Things like, hey, if we know that we need a cost center tag on every EC2 instance that spins up, let's spend some time in the early days of our journey on putting in some sort of script or automation that actually does that on our behalf, right? Or alerts engineers when they might have forgotten things of that nature. The earlier you do that, it helps create that culture. And so, you know, an example that I always give folks is a mistake that I made early in my career as a practitioner.
I kind of started the FinOps program.
Like most folks, I see all of these opportunities for waste. I start just like a kid in a candy store. I feel like I'm on top of the world. I'm showing savings, I'm doing all these things. And there was a particular sandbox account where it was old, hadn't been used for years. I discovered it. I found like over the course of the year we wasted basically $500,000 on these old RDS databases that I had just been sitting in a sandbox account, right? Cleaned it up. I'm a hero.
Well, what happened is six, nine months later, a similar situation happened where that a different sandbox, different team, but same kind of core structure of just these idle RDS instances were just left in the sandbox.
And in that moment, it's kind of this realization of if I had taken some time to either create some alerting or maybe because this is a sandbox account, like putting in some lights off, lights on type of automation. If I had just spent some hours early on doing that, well, we prevent this waste from ever coming back, right? And I think that part is easy to overlook because there's so much to do when you're starting out in finops. But that kind of taking a step back, thinking strategically and thinking proactively is really important. And what kind of I try my best to stress to folks now.
[00:23:31] Speaker A: Yeah, no, I think you raised a good point. I haven't thought about that. Like, you know, you sometimes, like you have the pressure sometimes maybe for your, you know, your objectives or your management that you need to clean up things. Especially at the beginning. Like at the beginning it's all like, you know, you are cleaning the cleaning department and you need to clean up everything so that you need to save, right? But like doing a cleanup without any communication or without any sharing is not going to help anyone in the org, apart from you and the org money, of course, but it's not going to, as mentioned, not going to help that from repeating itself on the future, right? It's something that is going to happen over and over again. And unless you govern it and you make some evangelization and some information to the teams that are originating that spend is not going to change. Right?
[00:24:29] Speaker B: And, and I think this is why I'm excited that unit Economics has become such a hot topic.
You know, the foundation has done a good job of pushing that and I hear folks asking us more and more, well, about unit economics. What can we do? How can you help us? The reason that's really important is because if I'm, if I'm, you know, let's say tomorrow I decide I'm going back to the practitioner world and I'm going to leave the vendor side and I hire on a new job, well, the pressure on me to show my value typically is around finding those big opportunities, right? Finding those $500,000 of wasted databases, that's typically where I can try to justify why I need to be hired in this role. But with unit Economics, you now can start to root your spend in something bigger, right? These kind of KPIs, that if I then start preventing those waste, you know, those databases in that sandbox account, and then over time as revenue goes up and I've prevented that from happening again, well, now my KPI start to look good and that's how I can root my value and that's how I can prove that I'm bringing something to the organization and, and all of that. And so I do, I hope. I don't think we're quite there yet, to be honest. I think unit economics still seems like a, a big scary topic to people and they overthink it to be honest around like what that means. It's just rooting something of value in your business, comparing it to your cloud spend. Right. That's, that's all it is and so we'll get better with it. But I think that conversation is going to help folks not feel so much pressure to just try to find all the savings they can day one, which would allow them to spend a little bit more of their time focusing on governance and focusing on policy for the long term.
[00:26:13] Speaker A: Yeah, you wrote a great point on the unit economics. I think we are mostly, if anyone has worked, especially in a big company, they are mostly evaluated through KPIs or through target metrics or through anything that means something to the business. And normally you need to measure that ways to do business KPIs that after all mean something to a business.
The problem, I think one of the problems that unit economic has is.
Well, it's not a problem is the nature of the thing is that this depends on the business. So like each business is very difficult to have like a set of unit economics that is going to work for everyone. That is going to mean a lot. Like you can have cost per cpu, but that's a unit economic. That on itself is not super meaningful. But for example cost per, I don't know reservation or cost per, I don't know per user that that one is the most generic one that can be useful.
But you know, it depends. Like meta will have one set of KPIs and then I don't know, Walmart will have another set of KPIs and that's one of the things that I don't know if you see it with the clients, but it's very difficult. Like the idea is pretty clear to me. But then the devil of the implementation, like unit economics is probably one of the things that is the most difficult to implement like in the day to day because it needs to be super specific. What do you think about that?
[00:27:41] Speaker B: Yes and no. I do think that perhaps the most valuable or more granular you can get with like, you know, I go back to I mentioned earlier like cost per transaction for a healthcare company or even a banking company. Right. Something, something like that, that, that helps you get a little bit more insight especially into particular applications or products.
But I think if you're Just trying to get started. There are some easy ones that are not org specific.
I mean, even as something maybe you could argue if this is a really a unit or not, but in my opinion, it's still valuable. Like just cost per revenue. Right. Kind of like what's your margin?
[00:28:19] Speaker A: Right.
[00:28:19] Speaker B: Is your, is your revenue and your cost staying flat? Is your revenue going up a little bit faster than your cost is going up? You know, things like that. Just to get yourself rooted as an org something everybody understands how much revenue is the business making and how much is our spend.
[00:28:36] Speaker A: Right.
[00:28:36] Speaker B: I think that helps people get started. And then you start to figure out, are there more micro metrics that my team or my application could take on that tie in to this number. Right. But are a little bit more granular for us to manage.
[00:28:52] Speaker A: Yeah, I agree on that.
I think we have an initial set of things that we may or may not. And you need to start with the easy ones because, for example, I can tell my own experience that I normally try to when I do in a metric. I want to create the biggest, more complex formula that measures all the factors in a single number. That is the definitive metric. And all of that and maybe with two, three metrics unrelated, not correlated to each other, but they are meaningful, like health indicators.
You can already do that and it shouldn't be as difficult especially. And with everything it needs to evolve. Like, you cannot start trying to make the biggest, more difficult calculation of analytics while you don't know how to do that. Like, you don't know how to obtain that data. So you need to start with, okay, like few information or like not super relevant. As you mentioned, you know, revenue per. Or related to cost to cloud goes is something that is, it can be more or less useful depending on the company, but it's something. And then from there you can do revenue per revenue. So revenue per cost on each specific line of business. Okay. You're starting to see which ones are correlated to the cost, which ones it isn't. You know, you divide it and then you, you start to evolve by doing and not thinking about, okay, I need to do the perfect unit economics for. From the first side. Right. What do you think about the, the evolution?
[00:30:30] Speaker B: I think you nailed it. Like, you start to like kind of start at the top and then things will start to make more sense for each individual line of business. I, I also too, I know, I know we're kind of straying away from policy and governance, but one other quick plug that I would tell folks here is the other thing and the foundation has done a great job over this past year and things are taking off this year and will continue to. But the beyond cloud visibility. Right as the as the foundation calls it scopes folks need to really put an emphasis on understanding what those material other costs may be SaaS, licenses, maybe it's data center. And before you spend too much time on your what is our perfect unit of measurement, you need to have some visibility into that spend because without it, especially for organizations that have millions in spend in other items, you're only looking at a piece of the piece and you're not actually taking in total cost of ownership to run an application. And so I would just caveat there or add that in to say for folks who are interested in getting started with unit economics. I would encourage and stress to you the importance of having more than just the cloud spend for an application factored in to whatever it is you're trying to come up with in terms of how you want to measure your KPIs going forward.
[00:31:48] Speaker A: Yeah, definitely. And this is something that has been brought up by another conversations because like we think the ones that come from cloud and we've been always on cloud, we think it's the major span and is the centric. But if you see it's not the majority percentage of the IT cost right now because you don't need to forget the licensing and the SaaS and all of that and the data centers because there is a lot of companies that need to or are still on data centers because of the nature of, of the business. And we've been talking about a lot of things and we've talked a lot of initiatives and how to handle incentives. But what's the approach that you have to enforce these initiatives, these policies that create some accountability to these teams without creating the friction of being the Phoenix police?
[00:32:46] Speaker B: Yeah, it's really important and it's also can be really hard to do. I think it's kind of a multi step process in my mind.
The first is you have to actually create policy and that means writing something down or typing something in a document. And before you do that it has to be collaborative. This is not. Well the FinOps team comes together and they say well these are the ways that we can save money. So let's create policies around this. It has to be finops perspective. The engineering team has to have a perspective and a seat at the table on how this would work for them or what makes sense for them in terms of their processes and then you know, leadership and other stakeholders. And so having a Collaborative discussion around targeted policies and then what that means to each stakeholder is the first step. And then as you kind of move through that, that process, you then start to figure out, well, how do I make the actions off of these now policies that become guardrails? How do I make this non invasive or as less least invasive as possible to our engineering teams? And that means meeting them where they are. If they use Slack, well then your actions need to be slack notifications. If they use Jira for a ticketing service, then your actions need to be a JIRA ticketing service. They don't need to go somewhere else to do finops, they need to go where they already go as part of their normal day job to continue to both develop and innovate and also take actions on, you know, finops related actions. And so the last piece too that I would call out here is that part of that collaborative policy conversation you should be identifying areas where you don't actually need to have a human in the loop. You don't need to bother an engineer.
If everyone at the table agrees that our sandbox environments can be turned off on nights and weekends, that doesn't need to be a notification for someone to go take an action. The next step there is we have that policy now. We put in automation to do that. And so finding those places where it needs to be low touch but drive savings and value and then balancing that with the things that do need action or do need approval or need some sort of human in the loop interaction, those things are meeting people where they are not asking them to go to a different place and things of that nature. So I think what you'll see too is as your organization starts to really take policy and governance seriously and if you make that collaborative enough, folks are then aware of what those policies are. And so a simple example is let's say that your company has decided we don't have a use case right now to justify using Anthropic's most expensive AI model. Doesn't really make sense for us. And so we're going to have a policy that says we're going to block that from being used.
If engineering had a seat at the table, they already know that policy now exists.
I didn't create it in a back office and then someone tries to go use it and it's blocked. Right. They already are part of that process. And so you're not disrupting them because they're already aware of it. Right. And so that's kind of generally how I think about the process of policy collaboratively. Then you start to put in those actions and then you put in real automation with no human in the loop where applicable.
[00:35:55] Speaker A: Oh, that's a great advice. I think you nailed the way that phenoms needs to be, you know, adapted to the situation of the company. The element, the elements that they use, like the slack, the jira, you know, it needs to adapt and evolve and not force people to do certain stuff because people are not going to like be able to, let's say do whatever you say because you say it and they are going to, you know, try to make it as easy their lives easy as possible while helping the company. Right, that's the way.
And wanted to ask you because we've been talking for more than 30 minutes and we don't know a lot about you and maybe you can explain a bit about your story and what do you do at Kion? Maybe some briefly mention about the story that you guys have.
[00:36:48] Speaker B: Yeah, so I'll do a quick overall story and then talk a little bit about Kion specifically. So for myself I mentioned earlier I was a practitioner in prior life. I basically like many folks stumbled into finops. There was a need, we didn't know how to allocate, spend, cost out of control. I kind of raised my hand, got involved almost 10 years ago now and then kind of made a career in FinOps. The foundation then comes out. I finally knew what I was doing for a living. I could point to that as like oh, I'm doing finops, right. I think I always joke and tell people my title before 2019 was R D effectiveness because we didn't know what else to call what I was doing. But so I did that at a large healthcare company for many years and then at FinOps X in 23 I met some folks at Kion. I was kind of looking for a change. I thought about maybe going to the vendor side. I was a user of, of one of the tools in the market from the very beginning and so I kind of grew up with tooling. I knew what I liked, what I didn't like, what I felt was missing.
And so met some folks at Kion and then a couple months later came on board to be, you know, over the kind of finops products at Kion. And so now two and a half years I guess in of being here and a little bit about kind of the product is, you know we're really focused on automated governance for finops that that's our niche. You know we do traditional reporting and dashboarding and metrics and we kind of COVID you know, those basic, what I call traditional finops use cases but where we are different and where I spend most of my energy and my time is related to how do you automate policy and governance for organizations. And now with the rise of Magentic AI, like what is the role that plays into that? And so between our automation and investing in our in app agent experience, that's where a lot of our investment is going to right now and where we're finding, you know, success with folks who are feeling like they're getting real value and something different out of their, their finops tooling. And so that's kind of what Kion does and what we focus on.
[00:38:56] Speaker A: Okay, no, that's a great mission and I think that you, you followed like a normal path and you know we get like very experienced people coming from like the end user now. They were no like that they were doing freedoms and then you know, because of dealing with all the freedom stuff they end up with the vendors and you know, improving the develop, you know, just to, to continue is like if someone is starting like you know yourself like 10 years ago, starting on the phenips journey and if you want to give them some, some advice or want a piece of advice, what it would be.
[00:39:33] Speaker B: Yeah, I, I have something here that I don't think is unique to me, but it definitely tailors to my experience and some of the roadblocks that I had early on, which is I think in many organizations, especially large organizations, you often have leaders or teams that either had a bad experience with FinOps at a prior life or don't quite get what it's ultimately supposed to be. Right. You might have an engineering leader who just sees finops as a roadblock for their engineers. Right. Things of that nature. And so what I would stress to folks who are starting out is don't worry about labeling anything yourself. The processes you're doing. Just start doing things that are valuable. Start collaborating with engineers, start setting up business reviews that are highlighting the wins that the engineering organization is doing. I think that's a big one is like if you start showing people hey and giving people maybe more credit than they even deserve to be honest. Like if you maybe you led the initiative but someone played a small role, make it seem like that person played the biggest role possible. Because what you start to do is that if you don't label everything you're doing as finops, it's just trying to make more valuable processes and save some money along the way, next thing you know, you've properly kind of highlighted wins. You've really started to create processes that are creating value that you can measure and show.
Then you can start to talk about all the things I just did are finops.
[00:40:58] Speaker A: Right.
[00:40:58] Speaker B: But at that point you've already won people over. People are bought in, they see the value you're providing. And I just think sometimes if you come in and everything you're doing is FinOps business review, FinOps tool, FinOps this and that. It can have a stigma that hurts you more than it helps you in the early days. And so just come in, figure out how to create value and worry about all that stuff later is I think my biggest piece of advice for someone.
[00:41:22] Speaker A: Starting out that's very interesting and I haven't thought about it, but after all, the way that we are introducing the finops is we do things that are oriented to the cloud cost optimization or the IT cost optimization and then we understand that it's called FinOps, but not before it's not. First phenoms and then, okay, then that's what it does. Normally you are introducing to the topic and then there is a whole framework, there's a whole ecosystem on that. So I think that's a wonderful piece of advice, especially for people that are in the very beginning of discovering finops. So thanks, thanks for that and you know, I think it's a great way to end up the conversation. It was a pleasure talking to you. I think there is a lot of value on this.
Thanks for taking the time and being here today.
[00:42:16] Speaker B: Yeah, I really appreciate it. Thank you for having me.
[00:42:19] Speaker A: Yeah, pleasure. Take them and see everyone in the next episode. Bye. Bye.
