← AI to ROI
Loading summary
Transcript38 lines
- [00:00]
A
Foreign. Welcome to the AI to ROI podcast, the Big Story episode. I am Ray Reich, I'm the founder and CEO of BenchmarkIT.
- [00:16]
B
And I am Peter Buchanan. I am the managing partner at New Plan.
- [00:21]
A
And this is the weekly episode where we dive into the AI OI Newsletter's big story of the week. And you can find that on substack@ai2roi.substack.com Peter, can you introduce this week's big story that we will be discussing today?
- [00:39]
B
I can. So this week we're going to examine three SAS to AI first playbooks that are used by market leading companies that have made that transition. But before we get there and we talk about that companies. Ray, what does it mean to be SaaS? SaaS to AI First Company and how do they get there?
- [01:01]
A
I'm going to do it at a high level first Peter. And it's those SaaS companies that aggressively transform their products, their product architecture, their GTM strategy, pricing and culture to reflect that of an AI native company while leveraging all their pre existing motion advantages. This isn't a simple plastic surgery or putting lipstick on a pig. This is more like an organization replacement. They are going to act, think and execute like an AI native company.
- [01:28]
B
And why is it urgent now for SAS companies? I mean things look great for them a year ago, but now they seem to need to go as quickly as possible.
- [01:39]
A
Yeah, one word. The SaaS Apocalypse Now. This was kind of stimulated by Anthropic's release of Claude Codework on January 30th and investors freaked out. In fact, from February 3rd to 6th, just in those like three days, over 300 billion of market cap evaporated from enterprise software evaluations. Over a trillion dollars evaporated over a month. And even the biggest and the best big size companies were not immune to this. Salesforce down 46%. ServiceNow down 54%. Hubs down down 51%. So the market is saying this is an imperative and it's a right now imperative.
- [02:24]
B
Wow. Well, all right, well let's step back from that for, for just a minute. SAS companies aren't going to disappear overnight. So they, so this seems more like the battle for the soul of the software industry over the next 24 to 36 months. Who are going to be, who are really going to be the surviving company? So Ray, how would you frame that battle?
- [02:47]
A
Well, I think there's three kind of categories broadly defined. There's legacy SaaS, companies that don't evolve, we can just put them to the side because they're not going to exist in two or three years. Then there's a saskpocalypse impacted companies, they're the ones that have a great system of record distribution data to their advantages. And they are going to become an AI first company by doing these five things. They're going to build and architect AI native products from the ground up, not just throwing generative AI or a prompt engine on their existing SaaS product. They're going to deploy an agentic AI layer that automates workflows while leveraging context. They're going to change your pricing to be more focused on either the work conducted or the outcomes delivered. Their GTM strategy is going to be AI first and they're going to have this AI first culture where no process, no hire, no org structure, it remains immune to being re engineered. For me, an AI first company or quite frankly, like in AI native companies, you could say that AI native newcomers would be the winners. And I'm going to quote the great Lee Corso. Not so fast, my friends. Legacy SaaS who do those things have a real unfair advantage.
- [04:08]
B
That's right. So today we're going to do a deep dive into three companies that have actually made this transition and are just knocking the ball out of the park. So you want to know who the companies are, Ray, that would be helpful
- [04:21]
A
if I need to talk about it.
- [04:22]
B
Peter would. So the first is Notion, the productivity suite for teams. The second is Canva, the really superior platform where pretty much anybody can turn into a creative and create great creative products. And the third is the big enterprise software company ServiceNow. They've all taken different approaches. So let's get started with Notion. Notion was a successful SaaS company long before OpenAI entered the chat. They've always been about collaboration, but now they're all about agentic AI. So how do they do that?
- [05:06]
A
Well, I'll quote Notion's founder, Even Zhao. He calls it agentic execution. And this is where Notion AI doesn't just assist users, it actually completes the work automatically on their behalf. And it's interesting product evolution that really transformed their existing SaaS product through some acquisitions. They, they actually acquired Automate IO in 2021 and it became the foundation layer for workflow automation. Then they acquired Cron, which became the Notion calendar. And then they acquired skiff in 2024 and that really enabled Notion Mail. So each of these acquisitions added additional capabilities and provided the foundation for more agentic AI execution.
- [05:54]
B
Yeah, so that led to some huge upgrades to the platform in the last less than a year. There have been two major relations. It's a releases, it's a completely different product for teams. So what happened with those releases? I mean, it's just a completely different environment, right? Completely different.
- [06:13]
A
The lotion deer 3.0 product, which is powered by the Notion Agent and it goes beyond kind of that traditional retrieving information for collaboration and it builds databases, it creates forms, it manages the end to end projects and it executes a multi step workflow while users can focus their time elsewhere. So it goes to execution and kind of for the founder, that agentic execution. And then in January they added more AI features. They added mobile AI notes with realtranscriptions. Then they added the multi model selection. So based upon the right model for the right task, it can use GPT5 from OpenAI or Cloud Opus 4.5 from Anthropic or Gemini 3 from Google. And as we've said again and again, it does provide that autonomous background execution by the AI agents.
- [07:07]
B
Yeah, so they also changed their pricing. So they used to be a product that was eight to ten bucks per user. They didn't really have a freemium tier. When they decided to go to AI, they didn't want to have an upcharge and get a whole bunch of customers sort of stuck at the lower tier. So they, they basically doubled their price. But because they'd done these acquisitions, they had added all of this particular functionality. Their users looked at it and said, well if I went separately to use all these functions they put in here that I like, I'm, I'm actually going to pay a lot more money. And it also changed adding this AI also changed their go to market significantly. So what do they do there?
- [07:56]
A
Well, Notion's distribution strategy really exploits a critical insight which is that knowledge workers aren't always fans of their enterprise software. So Notion used a traditional product led growth go to market strategy that built these viral referral loops through free individual plants. So it allowed them to sign up, have their individual plans and collaborate with their fellow employees. So by the time the corporate IT department discovered Notion employees had built it into their daily workflows. So most CIOs, you know, they take a hard look and say, hey, do I want to stick with existing SaaS applications that apparently my people aren't using and go with these AI first products like the Notion agent. So they really leverage PLG to then get enterprise license with their enterprise customers. And I think it's paid off financially pretty well. Right Peter?
- [08:49]
B
Yeah. So you know, in 2022, based on an analysis from Saster, they had revenues of $67 million. Very credible, profitable company. They doubled from the year before. But if you fast forward three years, they are projected to have ended up 2025 at $600 million in revenue. So in three years that is a 9x revenue growth. And it's because they're monetizing value at a higher level, they're adding more features. Basically people are. The application is just stickier and it's harder for people to leave it. Which means Ray, that they've got some really fantastic moats AI native type moats.
- [09:36]
A
Well, I would say that 9x growth over 3 plus years is pretty incredible. And I think their acquisitions and building agentic AI from the ground up, I think that was really critical to become an AI first company. But those more legacy mostly have, you know, they include things like the platform itself creates the data network effects that's hard for competitors to match. Users 1 gets users 2 and 3. Users 2 and 3 get users 4, 5, 6 and 7. Every document collaborated on every conversation can create context that makes those AI agents even more effective. The switching costs aren't just about migration hassle. They're about trying to rebuild the organizational knowledge that now is embedded with the Notion that makes the AI agents even more powerful. And as teams customize their databases and workflows, they're effectively building proprietary software on Notion's agentic AI infrastructure. So all those things made them much harder to displace.
- [10:40]
B
Wow. Well, they're in great shape, so let's talk canva. So a good way to describe their mission. They want to democratize design through AI powered communication tools. So whatever you want to build, they want you to use their platform and their AI features to create artifacts that you can use as part of your business presentations, worksheets, data sheets, business cards, logos, whatever you want to do. And they've also been around a long time, so what was their path?
- [11:18]
A
G' day mate. Let's get under that. So now they're from down under. So sorry to the listening audience if Maya's drawing accent wasn't very good. But they had several kind of advantages. Number one, their offerings were already beloved by millions, including yours truly when AI came along. So they took their existing product which allowed someone like me who was a non designer to be a designer. And they supercharged it with AI to make me an even better designer and to expand into near adjacent markets. And they didn't really just bolt AI onto their existing design tools. They rebuilt some core workflows around AI agents and assistants. So now the platform offers AI native features, such as an AI powered background removal, Magic Write for copy generation, Magic Edit for photo manipulation. They have a text to image generation tool. And they all do this without something that I don't have. And that's design expertise. They turn non designers into AI empowered designers.
- [12:24]
B
Right. And it's basically those AI features are all invisible. They just go to work for the users to make them better.
- [12:32]
A
Yeah. And the other thing is they've even said, okay, we know which LLM model is best for some specific general capabilities. So they don't make their customers need to decide which LLM tool you use. It's built in based upon the process or task they're trying to execute.
- [12:50]
B
Right. And the other thing that's been really helpful for them is they have like notion multiple pricing tiers. They have a very pretty powerful freemium tier. So if you just have to create something every couple months and you want access to some templates and you want a little bit of AI to help you do it, you can actually do that. And those freemium tiers are good enough that it sort of whets the appetite of users that kind of want to move up the stack, Right?
- [13:21]
A
Yeah. Well, we talked about the pricing, we talked a little bit about all their AI native capabilities integrated. How has their financial performance kind of reflected these investments they've been making?
- [13:32]
B
Well, so first of all, they were already a big company in 2020, they were half a billion dollars. And so they came through the pandemic basically like a house on fire. And if you Fast forward to 2025, they're a private company also, but they'll probably end the company based on what's available, the data that's available in the press, $3.5 billion. And they also have these tremendous, tremendous usability metrics. So as of August 2025, they had 204, 40 million monthly active users.
- [14:10]
A
Wow, that's up like, isn't it up like 33, 34% year over year?
- [14:15]
B
Yeah, it's up. It's an incredible change. They have 800 million uses of their AI tools every month, which is up 700% from last year. And they have users in 190 countries and there's only a little over 200 countries in the world. So they're pretty much everywhere. And so they just provide incredible value. There are other, you know, creative tools like Adobe has their suite, Figma has their suite. But from a price value perspective, this is really fantastic. So if we go and we talk about Moat, okay, they have Moats they have the same sort of viral network effect brought on by their premium tier, brought on by just the fans that they have everywhere. They have content lock in because they have hundreds of thousands of templates that, that companies can use and examples and communities. They have the sort of invisible AI personalization that makes things better and they have the ability to sort of walk people up the stack and it's reflected in their growth. It's just been really, really tremendous. We love metrics at AI to ROI and we love moats and they get a plus on both.
- [15:45]
A
Yeah, just to talk a little bit about something that you mentioned and that is their actual uses per month, right? 800 million. It really shows you how some of these new AI features have really increased the stickiness in the utilization rates. That's a 700% year over year change as far as their AI tool usage. And I just wanted to reiterate real quick that that template marketplace, you know, I remember many years ago in a marketing automation, I created this concept called Blueprints where I provided templates for different marketing campaigns. But Canva takes it a lot further where that template marketplace actually allows users to generate their own templates which attracts more users and actually their users are helping create the product. So I'm really excited about Canva, but there's another enterprise software company that I'm really impressed and I think if there's going to be one really big winner, I would bet on this one and that's ServiceNow. Can you talk a little bit about them, Peter?
- [16:44]
B
Yeah. So ServiceNow, their approach to AI first is different because they're a really big enterprise company. Their revenues are in the multi billions, they're much bigger than Canva and so they're famous for low code, no code development. But they're also famous for helping companies manage workflows on an enterprise scale. And so what's the current state of service now from a technical stack perspective? Because they've changed a lot in the last five years.
- [17:21]
A
Well, you mentioned they've spent almost 20 years building this workflow automation infrastructure that now AI agents can orchestrate. So the platforms kind of now assist AI. They don't replace what was there. They enhance that workflow layer by enabling AI agents to execute and orchestrate workflows autonomously versus routing those requests with human operators. So if step three was saying hey, get this human to review and approve, that's now done by these intelligent agents. So that agentic orchid across cross functional processes and multiple points of system integration provides a real advantage to them.
- [18:06]
B
Yeah, and they, they've been on the M and A trail for the last five years too. So they've, they fought 22 companies to add capabilities to their platform since 2022. There are, yeah, it's a lot and, but there are four that are pretty recent. Two are, one is Armis. They're paying, their transaction is not quite closed yet. They're paying $7.9 billion for it. And it provides AI powered cybersecurity capabilities. And then there's another cybersecurity company called Visa that they bought early in 2025. And they're basically honing in on identity management, asset discovery and IoT devices, which were holes in their platform because they want to have, I mean you can create basically anything you want inside ServiceNow. So they need to have an infrastructure stack that goes with that. They bought Moveworks and Q&AI to provide agentic capabilities and they bought another company called Data World that concentrates on data governance. So you can see they're building an environment around those agents and those workflows so they're successful and secure and also so that the data data that customers are working with is high quality data because that's the, that's the number one factor in the success of agentic AI. So they've also done model integrations and spent a lot of money with OpenAI and Anthropic. They've committed $200 million each over the next few years to have those models integrated into their platform. So you know, they have a really, really kind of a big strategy here. So let's talk about their pricing because they have to make this same AI first shift in the way that they price to price not based on seats, but based on other factors. So what are they doing? It's a big swift.
- [20:16]
A
So if you think about this kind of cross function business process workflows that they automated, it was still human enabled. That means humans kicked off a process, humans reviewed a process and moved it to the next step. Now because the agents are doing a lot of this, they've moved from that seat based pricing where humans are involved to task pricing. So it's where customers pay per an automated workflow executed. And this is really looking at the end to end process. And this transition actually acknowledges that the agentic AI layer that they are now layering in actually is going to reduce the number of humans while increasing the workflow volume in the outcome. So their pricing is going from human based to task based. And also I think that they have modified their go to market strategy beyond Pricing correct?
- [21:13]
B
Peter they have modified it. They've had to do a lot of retraining of their sales force so that they're not selling seats, they're selling outcomes, they're selling these tasks. They're selling more of a partnership with the customers. And Bill McDermott, their CEO said they've had to just retrain their sales force completely in the way that they actually talk to customers. Now, if you've ever heard Bill Dermot McDermott interviewed, you totally want to buy something from him 30 seconds after he starts talking. So I imagine that went down, that flowed downhill into the sales force and they've been successful with that selling transition.
- [22:06]
A
I'll tell you, Peter, what's interesting because you know, Bill McDermott came from SAP, so even though he's really innovative with where they're taking the product to become an AI first product, they also don't take their eyes off the ball for financial performance. So, you know, even though their stock got hit, as we mentioned earlier in the episode, that was really throwing the baby out with the bathwater. Because if we look at their financial performance both over the last five years, but even over the last year, I mean they have a annual growth rate of over 21.3%. Their EBITDA margins are in at 26, 27%. So that says that their rule of 40 is around 47 to 48%. And even in the last year, if I'm not mistaken, I think their growth was around 21%. So they have done all these 22 acquisitions, migrating from a legacy SaaS to an AI first software company and they're still knocking it out of the park. And one of the reasons they're knocking it out of the park is because of those most we talked about that they leverage while they become an AI first company. The fact that they've already passed a lot of the data security and compliance demands that enterprise customers have, so they've got a lock in there. They've got deep integration across sometimes hundreds of systems. It's hard to rip out the workflows because they're end to end, have a lot of complexity. So the switching costs are very high. And now they're going to enable companies to reduce the human headcount or at least have humans focus on higher value activities because of that orchestration layer and continuously building the institutional knowledge. So I mean, I am so excited about ServiceNow.
- [23:49]
B
Peter yeah, there are reasons to be excited. I think of the large SaaS companies, they're probably the one that's most likely to survive and thrive without major hiccups in this environment. I mean, so let's sort of end here with some common themes. So there are three principles that seem to determine the AI first winners. So the first is pricing has to align with the value that AI creates. So per seat models don't do that. You know, AI first models, they price workspace access, they price workflow automation, they price outcomes. It's more of a partnership with the customer. I think part two is the product architecture has to enable AI agents, not just assist humans. So features aren't the thing. The product architecture has to enable outcomes. And so the architecture has to have APIs that has lots of integrations, it has very solid AI infrastructure and the agents have to operate autonomously to support those outcomes. And then the last piece is the go to market strategies have to support AI economics. So basically, bottom up adoption works because AI makes the products valuable to small teams. You get viral loops, you get more people using the products. That's the Canva notion playbook. And then you get the, in the bigger companies, you get these large projects, these outcomes. If we circle back to ServiceNow for a minute, Agentic AI is going to make these workflows that they've spent 20 years building more complex and valuable, not less. Right, so. Right. So, so how would you, how would what, what's your, what are your, what are your takeaways? You know, sort of as, as we. We close up this episode of the
- [26:04]
A
podcast, I'm going to wrap up with one thing. So the SaaS apocalypse, that wasn't a critique on business fundamentals, it was on market valuation fundamentals. And I believe that those SaaS companies that are going to emerge to be even stronger are those that implement these AI first variables that we talked about. And they need to rebuild not only their business model, but their business culture and ethos around the fundamental truth that AI is the future of software. Agentic AI is the future of getting work done. And companies that embrace that and highlight it are going to be the winners. So, Peter, hey, thanks a lot. This is a fun episode.
- [26:48]
B
Absolutely.