How To Prove AI ROI to Your CFO

If you're a tech executive that's trying to build a business case and really understand what's the ROI for moving from a software development lifecycle to an AI DLC or AI development lifecycle, take a look at this. We built this little calculator of what's the cost of this, what the returns are, and I'm gonna explain that real quick. The real things that really compound are we unlock like the ability for people to do more things after automating and augmenting certain workflows. And that increases your best decision maker's intellectual capacity to go drive more decisions towards whatever. Might be product management to discover more opportunity, might be engineering to be able to increase feature velocity towards the right thing. And so really here, the thesis in terms of ROI is more people can become builders and contribute at a tech company. Like think about just PRs as a proxy, right? If you get more PRs and that is some proxy for feature velocity, not perfect. And you say, hey, there's a percentage of that that incrementally moves the needle on incremental market share capture or revenue. This is sort of how I modeled it is think about how big your team is. So this calculator, we've kind of had it at like 20 or below. So let's say we have an organization of like 500. And the thinking here is you could put in a fully loaded FTE cost of one of your builders. And right now, like in Silicon Valley, I typically see $250K. And then we say, hey, there's a certain amount of spend that we have to put on the CapEx side For total AI tech stack, right? So coding assistance, platform licenses, seats, API spend, etc. And then there's how many people are actually dedicated full-time to running this AI DLC transformation thing. That's something that I've seen at companies is unless it's like really CEO-driven, then this program team is just part-time people. And so you really want full-time people. And that's what this model has. So let's say you have like 10 program team members that are central driving sort of a federated adoption approach. And then there's this realization rate of how much more time are we freed up by automation and augmentation? And how much revenue do we start with? And how much more revenue can we capture through that velocity that I mentioned? So here, we just have like a 5% revenue capture conversion metric. There's another part of this equation. So there's like ROI, right? Investment is like the CapEx, the OpEx, the people, the tools, the technology, and then there's, okay, we can make money with that. There's a revenue generation aspect of it. And then there's, can we drive down cost? get more cost efficiencies. And one of the things I've heard from customers and prospects, especially with like hundreds or thousands of people in the technology or an R&D org, is duplicated effort. How many times do you sit in a meeting and then you're like, holy shit, I should have seen what this team has done months ago, because we're doing something similar and the Venn diagram is too close. And so we duplicated effort. That's a huge thing. And so you could kind of like forecast, okay, if I put in, this reduces some amount of duplicated effort, let's say 10%. What does that look like? And what's the breakeven point for this investment? And then headcount savings. So this is the OpEx part, right? Like a sensitive topic. But for everybody out there, you know, what's coming is, and we're seeing this in the news constantly, like layoffs and stuff like that. And I expect more in Q4 coming into Q1 next year. That's usually when these things happen after the teams have gotten some time to figure out who's really AI-pilled and who's on the bus and who's off the bus. Those are the conversations I'm hearing at the leadership level. So you could put this in there too for your CFO. They definitely want that number. Another use case that I've seen in particular for the enterprise story is, oh my gosh, Lovable or Replit or whatever, they make cool prototypes, but holy shit, what if we like took one of those and made like a little vibe coding tool for ourselves with our own design system? Can we arm our sales teams to be able to go to a prospect, they need more additive features, and then instead of going to a designer on the R&D team and 2 weeks later, then you get a prototype that looks in your design with your brand guidelines and then You close the deal. You could just arm that sales team to just go close on their own, and then the prototype they show to the customer could be in the design system that you have grounded in what the brand guidelines actually are. So then, if the sales team closes more, which we've seen, I just saw a story about this from one of our customers where they took a six million dollar lost opportunity with a prospect, and then within 24 hours, the CEO was able to, with their own tool, with their own harness, prototype something in their taste and design. System and brand guidelines. And they actually won the deal and then extended the ACV. So they won a $12 million deal. So it was a lost deal. They came back within 24 hours because of the speed of using these AI tools specific to their organization, specific to their own product and how it looks. And so this prototype ended up looking like exactly how it would look in the product, like not like a Lovable or a Replit or whatever with like different tokens and typography, you know, 'Cause getting that stuff is kind of hard, but they made their own investment. And so it's like, holy crap, if you are able to invest $500,000 into a tool to prototype for your sales team to close deals, and then you could turn around a loss, $6 million to a won $12 million, insane value there. It's a very easy thing for B2B sales-led organizations to look at as a tip of the spear for your own ROI case. I'd recommend looking at it this way and definitely sales-led orgs. If you're a product manager or an engineer at sales-led orgs and you're trying to Your ROI use case, look at enabling more sales. The CEO will love you, the CFO will love you, and the ROI case is going to be very simple. They're going to be like, "Oh wait, we invest more people or tooling technology costs into enabling sales to sell better and close more deals." Holy shit! That's like very easy to see on the top line. So we put in some numbers there, assuming that how many more deals can you win per year, and what's your average deal size, and what does that return? what's your win rate lift based on that? And then based on that story, like ACV lift as well. So turning around like something that was a loss into like a higher ACV win, that's what this model shows. And then down here we kind of have like some baseline and this is— there's a lot of assumptions here. Down here is all of our assumptions and all of that research. You could take a look at that, but there's some workflows in the traditional SDLC to look at and then prioritize. So if you're a product manager trying to build a use case, or if you're a program manager trying to drive transformation at your enterprise tech company and you're responsible for transforming the SDLC to an AIDLC, go look at the workflows that your designers, engineers, and product managers do. Here are some, and go put an economic number, like a financial number on it, and then stack rank it. And this should be, hey, can we drive down cycle time? And then can we do that and replicate that success across the organization, dozens of builders, Hundreds, thousands, and then get to an ROI that your CFO could really understand based on driving down cycle times. Just wanted to show this. And so based on this 500 engineering team and product size, here's like the bear, neutral, bull case. So bear case, just with the numbers I went over, is net benefit in terms of productivity, incremental revenue capture, cost efficiencies, deduplicating effort is about $5 million from a steady state point of view. This is really like, I think this is really important because when I talk to prospects and customers, when you go inside the organization, there's lots of people trying to figure out AI DLC. And usually what happens is, what I'm seeing is there are pockets of AI-pilled people in marketing and sales and product and design and engineering, and those AI-pilled people really get it. But then the rest of the org maybe doesn't, or they've tried these tools and then they have so many turns with the tools and it doesn't generate productivity, so they don't trust it. If you are struggling with driving change and championing change and getting people aligned, Really get down to what is the success metric in a way that your CFO can understand. And when you do this and you go through this exercise, I find that it becomes way more clear and cuts through the noise of any qualitative like, hey, I feel like we should do this. Hey, I feel like if we adopt more here, we'll get more productivity. It's like, what does that even mean? If you need resources, you're going to have to justify the ask to the CEO, the CFO, or the powers that be at your organization. And here's a little tool that you could use to do that.