Jason M. Lemkin and others discuss how enterprises budget for AI vendors. Once a vendor enters a company's system, budgets typically double yearly. A Morgan Stanley study found a 50/50 split between AI spending that replaces existing software and spending on entirely new categories, with much of it drawn from service budgets like consulting and legal outsourcing.
Chapters
a new vendor to put into the procurement system to get approved, to track. And every, and every enterprise knows that if I bring in one, I, it's not just one year for Synthesia. Now I'm stuck for 8 years. Like that's how enterprises think. They don't think one year. They're like, okay, it's gonna take me the first year I'm learning, the second year I'm rolling it out more, the third year I'm gonna buy more stuff. And they always like double. We at, back at Adobe, like when you put any vendor in the system, you doubled the budget because you knew it was gonna grow with add-ons and seats. Right? And it was, and so how many of those can you add each year? Even at Adobe size, you just run out of vendors you can met. And I don't know if AI change, if agents change that rule. They're gonna change a lot of other rules. And so I don't know who's gonna pay for that. I just don't know who's gonna pay for the discrete vendors, not the agents themselves, but so many vendors. So many vendors.
I do think they're like, there are fundamentally new product categories though. I would put us in one of like, we're never going to replace someone or take budget from someone. We probably will over time.
You don't come from the training budget or the LMS budget if you're for an internal, for an internal use case.
No, we'll, we'll sometimes take some of the video production budget that they'll go that you use with agencies.
Yeah.
But the key point is that people are not evaluating us against should we, should we switch out this thing with some easier, right? It's a net new thing. It's also why we're growing so fast. And I think there are new product categories like that. And then there'll be a lot of these agents or features or whatever that'll exist better within, you know, the platforms people already use. And, you know, there'll be kind of add-ons and extra modules, I think.
I believe you that you listen, you know, you're in all the deals.
I believe you that you're not directly stealing someone's creaky old budget. But what I— my limited experience tells me behind the scenes, you're stealing someone else's budget. They're canceling someone that they don't quite need that was on the bubble. And this is— I mean, I don't think this has changed since I worked at a Fortune 500 tech company. You get together each year. This is why budgets annually. And Tomas, you want to add that cool product, what are you going to cut? This is why so many things got cut.
Maybe some of the budget, maybe some of the budget does not come from software, right? It comes from like services-based things. So for, in our case, a lot of the budget we eat into is like video production agencies, right? That's not software vendors or translation budgets or it's kind of like more service-driven things. And I think especially if we talk about agents, right? That, that, that would probably be a lot of what that could eat into. So it's like you may add another software vendor, but maybe the budget doesn't come from like other software platforms, but it, but it comes from like services.
That's a lot of the hope on X is that we can turn humans into software budget, right?
Well, Morgan Stanley ran a study in Q3 I think, and they said 50% of the AI budget was cannibalizing existing software and 50% was new.
50/50. Yeah, that's probably right, right? I mean, they got the data, right?
Yeah. And so I bet Victor is right. Like, it's translation or other kinds of services, consulting, customization, legal, like all that stuff, like legal outsourcing. That's where the software budget's coming from.
