All right, thank you all for joining us for the tech portion of the day. I'm Thanos Machopoulos. I'm the Canadian technology analyst at BMO. Happy to kick things off with Tom Jenkins, Executive Chairman and Chief Strategy Officer of OpenText. Tom is the architect of OpenText's strategy, having been CEO from 1994 to 2005. I know this personally because I met Tom shortly after the company went public in 1996.
Yeah, that's fun.
Tom outlined how the company went public as the search engine powering Yahoo, but he didn't view that as a sustainable business model. He articulated a strategy to go become the leader in content management. Here we are all this time later with $5 billion in revenue and the leader in content management. Tom, to kick things off, why are we talking to you instead of a CEO today? What motivated the change? Was it a question of a different skill set required for the next stage of growth? Was it a question of you wanted better execution, maybe provide the rationale?
There was no difference of opinion with management and the board on the strategy. In fact, the company is following exactly the plan that was approved by the board. No, it was more on pace, on how quick to move on divesting some of the non-core assets. That's really what it was.
In terms of who the board is looking for for the next CEO, what are some of the characteristics? Are you casting a very wide net?
The search committee, I'm not on the search committee. The search committee is made up of brand new board members because we want to take a fresh approach. The head of the search committee is the former Chief Human Resources Officer of Hewlett Packard, Kristen Ludgate. She just retired from Hewlett Packard in Cupertino a couple of months ago. Also, Bob Howe, who's the Chief Financial Officer of Fiserv, and Goldie Haider, who's the CEO of Business Council of Canada. Really well-connected group. They'll run off and do that. I think they're biased. Just like any other CEO's changes, you tend to move your pendulum. After me, you know I'm an architectural engineer. We got John Shackleton, who was a solutions engineer, Mark, another architectural engineer. I think you'll see they're biased to go to someone that's a solutions engineering background, simply because of product life cycles.
Mark just refreshed with Aviator the entire product line of OpenText. I think they'll probably look for someone to go sell it.
Great. Yesterday, you provided some new segmented disclosure outlining the revenue breakdown and cloud breakdown for each of your main product pillars. Let's talk about that. One of my key takeaways was the content business, which is 40% of revenue, growing at 4% last year, 17% cloud growth. Let's drill into that. What's driving that 17% cloud growth and how sustainable is that?
First off, we did the segment analysis. Chadwick, our CFO, had to go back to EQ Bank after what happened with Andrew Moore. We had promised that we would do the segment analysis so that you could have a roadmap to see where we were going and how we were going to shape the portfolio. You now have the data that we have. At the board, this is our data reviewed by the auditors, et cetera. We try to get it to you as quickly as possible so that you can follow along as we go to remove some of the non-core assets. Now, content, as you're asking, and the growth of it, it's really part of how we see going forward with the company. In a word, content is a very important part of training Agentic AI.
That's really when we started on this journey many years ago, it wasn't called Agentic AI. It was more like predictive analytics and Bayesian mathematics and things like that. You needed content to be able to train. You'll see Content Server and that whole business unit play a core role, but that's not the core of the company in and of itself. There are actually three kinds of content that you use to train an AI. There's human-generated content, which is Content Server. In the last decade, machine-generated content is enormous and, quite frankly, dominates the content that we're creating in the world today. That's something called IT Operations Management, or ITOM. One aspect of ITOM is actually machine-to-machine. Think of a nanosecond trader that creates log files. Those log files become really important for training predictive analytics.
The third group that's really important in all this data is business networks. That's because it's the information traded between organizations. That business networks division is very important to training Agentic AI. You basically have three. You have content, you have ITOM, and you have business networks. You need to wrap it in security because all of that data has to be secure. It has to be secure at the edge. It's got to be secure at the server. That was the reason why we bought Micro Focus and the HP software catalog, because they had all those components. The problem with something like an HP, it had not only all the pieces we wanted, but things that you would sell to Best Buy or to Mom and Dad, you know, that kind of thing. Those are the pieces in the retail part that didn't make sense to fit.
There's nothing wrong with those business units. They're just not core to what we want to do going forward. Really, when you think of it, it's not just Content Server. It's those other units. I think maybe the way to think about Content Server and the growth is it grew that much despite the fact that the management team was distracted by six different business units. If they have one thing to do, then perhaps that'll grow faster. That's hard to say. We'll wait for the new management team to talk about what they believe. Certainly, double-digit growth in the cloud, which is the core part of this thesis going forward, speaks for itself.
That's why we wanted to do the segment analysis so you could all see what we're aiming towards and follow us as we do it, as we judge us by how we create those business unit divestitures and track us. The strategy committee that got announced at the same time as all this other stuff is made up of, again, veteran members of the board, no management, because management is providing the information. This is a capital asset allocation. The board's going to do it. We're lucky enough to have folks like Annette Rippert, who's the Managing Director, just retired of Accenture out of Washington in telco and tech. Folks like Fletcher Preben, who's the CIO of Cisco, and prior to that was CIO of IBM. We've got lots of real veteran people to do this. Measure our progress.
We've estimated it's somewhere between three quarters and a quarter and $1 billion out of the five and a quarter. You'll see us take 15% or 20%. That's consumer products, things like developer toolkits and what have you. There's actually about 30 or 40 different product lines within those business units. That's what you'll watch us do over this next cup. It'll probably be one a quarter because when you divest things, it takes a bit of time to be able to separate out. That's what you'll see over the next three, four quarters.
In terms of the segmented disclosure you provided yesterday, should we think of it as being the SMB cybersecurity and the DevOps businesses going away in their entirety? Might there be pieces of it that you end up keeping?
Yeah, the devil's in the details on that because some of the edge security we want for the corporations that we sell to. There is a bit of devil in the details with that. I'm trying to just generally give you an idea of where this will end up. You're right. You're right what you're saying.
Going back to the strong growth in content cloud, it was just have early signs of AI traction. Were there other factors like the SAP upgrade cycle kind of pulling you along, or anything specific that you would point to?
You have to think of it. You know, all of you, I'm sure, like I follow this in the media and all that stuff. You have to think of AI as three different buckets. Follow me, if you will, for a second. The AI that burst onto the scene is really generative AI. It's the stuff that takes about a billion parameters. It's called a large language model. It captured everybody's imagination because it could do natural language. I can tell you, back in the day when we were talking about search engines, we tried so hard to let people ask a question. We didn't have the mathematics and the compute power to do what is called a complex Boolean. That's why all of you, as you used OpenText or Google or AltaVista, you could only put one word in.
The great thing about what you saw with ChatGPT and Perplexity is that you could talk to it in human language. That was the amazing thing about that. If you think of the other pieces of AI, the media talks a lot about something called frontier models. That's about a trillion parameters. That's what's coming. That's sort of the science fiction thing where it's both utopian and dystopian. We're not involved in that. That's going to be something that nation-states do. On the opposite end, with only, let's call it, a million parameters, so way less than what you're seeing in the public space, is Agentic AI, where you're creating agents. That's where all the productivity gains are going to be for corporations like General Motors or Coca-Cola or BMO, et cetera. That's the sweet spot that we're focusing on.
If you segment AI into those three pieces, we're trying to do the stuff that you're training, oh, say you're Marriott Hotels and you want a chatbot to refer to a convention that wants a quote on rooms in a city sometime next year. They need all the information internal to their organization about rack rate, occupancy, and things like that. To do that, maybe the best way to think of this and why content server matters so much, you, as a CIO and as a practitioner, to train that bot. Think of your house. You have to go up to the attic and pull out a photo album before digital. You've got a photo album of photos and what have you. You've got to scan them in. You have to put that as part of your overall training content.
Or you have handwritten letters that you've got to do optical character recognition with. God forbid, you go to your basement and you've got your old PC and it's got WordPerfect in it or VisiCalc. If you're under 40 years old, you have no idea what that is. Yet, if you're Pfizer or you're Boeing, those are the documents that have some of the critical information about a Boeing 787. If you're a CIO, you've got to go get all this stuff. You have to put it into a modern server, and you've got to make that server so that it can be made available to dozens of large language models that you're going to be using. When I'm trying to give you a peek under the covers of all this, there's a lot of stuff that's got to happen.
All these large language models didn't exist in their current form three years ago or five years ago. Everything I've just described to you is what OpenText built in the last two years. It's called OpenText Aviator, and it does multi-cloud. You may want it inside your firewall, or you may want it up on a hyperscaler, or you may want us to manage it as a private cloud. We have to be ready for all of that. You may want Anthropic as your model, or you may want Cohere as your model. It depends on which model you believe is best for your industry. Finally, you may want it for an ERP application or a CRM application. If I'm giving you a headache, that's the headache that we've had facing us over the last two years. That's why you saw our R&D budget go up so much. We're done.
We've built it all. It's multimodal, it's multi-cloud, and it's multi-app. It's called OpenText Aviator, and that's what we're going to go sell. We've built it all, and that's what Content Server is sort of the first indicator of.
Tom, let's talk about the go-to-market. I hear you that you have the technology ready to go. When you're trying to get the attention of the CIO and there are a bunch of other vendors who are also making the case that they're the ones to go to for AI, how do you get that mind share? How do you feel about the go-to-market motion today? How can it be improved?
Actually, it's the reverse right now. We don't have enough people to service the demand that we have. I think you'll see us start making tuck-under acquisitions for some of the verticals so that we can service the demand. Quite frankly, this is the kind of thing where it's not just OpenText. It's the whole industry. We don't have enough people that are articulate in all those things that I just said to you. Actually, it's can we keep up to the demand? That's the reverse problem. Now, how long will that last? Who knows? The reality is right at this moment, every organization in the world is paying attention to this. Because if they don't, they'll be out of business. This whole idea of AI first, you've seen the tech companies do it primarily first. OpenText has got its margins up quite nicely by doing this.
I think you'll see all the CIOs of all the organizations have to do this. Our challenge is to keep up to it.
Another takeaway from the disclosure yesterday was that the analytics and IT Operations Management businesses had double-digit decline last year. What would be your outlook for those and the plan to stabilize them?
Yeah, on top of all the other things going on with OpenText, whether it's the extra R&D bump that we had to do with Micro Focus, the board, the management, we all thought it would take a year to integrate it because we've been doing it for 30 years. We've done well over 350 different acquisitions and mergers. Guess what? When you buy Hewlett Packard and it's $3 billion and it's got six different business units, it doesn't take one year. It takes two years. We didn't know that at the time. I think we obviously know it now. What you saw in that second year was two-year maintenance contracts. Sometimes they're firing us. Sometimes we're firing them because we wanted to get rid of poor margin business. It's a little bit of a mix of that still. You'll have a much clearer view this year where you'll see them.
You won't see big declines like that. You'll see single-digit growth and single-digit decline. The other thing people have asked me, by the way, is generally, the margins are all the same. When we sold the AMC unit, that was unique because it was a mainframe unit. Actually, most of the business units that we disclosed yesterday, we're trying to get the margin information to you as well. That's harder because you have to do allocations. The auditors and the audit committee have to go through it. I cause this here. He was saying to me earlier this morning, they'll try and do it by the end of the quarter. That's a hard one for them because they have to unpick. It's like putting Humpty Dumpty back together. They have to unpick a lot of things to be able to attribute margins to each BU.
Generally speaking, the way you might think of the margins is we're about 35%, plus or minus 5% depending on the unit. Some of the units, and Coz has this in a lot of his analysis, where we do a lot of managed services, they're lower. Some of the ones that are SaaS, which are very high margin, they're higher from the 35%. Generally, if you were to think about 90% of the business, it's within a ripple of 35%. What we will do, like we did with the divestiture with AMC, we will always return to norm. We'll always manage the business. The reason why you might see a quarter or two of deviation is that, say you sell a business and you've got 6,000 developers and 2,000 developers are going to the divestiture, you still have the building.
You have to manage through either moving people or cutting the building down. There is a little bit of variation there. I think generally, you might think of it as a 35% business and it should be about mid-single digit once we get down to the core, which is what you're seeing with Content Server. Content Server is sort of a good proxy of what is our North Star.
I'll pause. Any questions from the audience?
That's interesting. We were talking earlier. The TAM of Aviator is actually going to be quite modest because it's not where we'll make the money. The same thing happened to us with search engines. We used to think that we would sell search engines. Search engines became an enabler. Aviator is really an enabler that connects a large language model to our content server. What happens is there's two ways we make money. When you're up in your attic and you're bringing down WordPerfect off of a hard drive or a mag tape or whatever, first, you have to convert it. We're getting a lot of conversion sales. Then you've got to reload it onto a content server because usually, when you have an archive, you don't pay anything for it because you just put it on a brick or a mag tape and you leave it.
Actually, Aviator is selling. I think you'll see from Aviator tens of millions of dollars. It will unlock hundreds of millions of dollars with content server, business networks, et cetera. It's a bit counterintuitive. You have to have it, though. You have to have it. It's like an API. You have to have it.
Tom, on the divestitures, I know that the process is being spearheaded by the board. Divesting a business is obviously a messy process. How do you ensure that it doesn't become a distraction for the management team just operationally over the next year as you carve out those assets?
The management team is actually a management team that's really oriented to selling to the CIO at General Motors. They're not a management team oriented to doing dealer networks within Best Buy and attach onto a laptop. It's actually those units are good units. They're just being ignored because it's not where the management team really has its expertise. I think you'll find that those units won't grow, they won't decline, they'll just stay as they are. Our initial indication is a lot of interest because there's nothing wrong with these business units. They just don't fit with what we're trying to do. There's actually quite a lot of interest for them since we made the announcement.
It meant just all the G&A heavy lifting that needs to be done to kind of split them off. I mean, are there separate teams managing that? How do you ensure that?
I was talking to Coz about that. You can ask Coz that. He gets all the hard questions. Don't ask me the hard questions. Coz just went through this with AMC when they sold the mainframe unit. They know how to do it. One of the reasons why you'll see us do this over many quarters is, we could sell all those units in one quarter. The dilemma is, as you're getting at, you have to do that facilities rationalization. It's better to do one business unit a quarter over the next two or three quarters. It comes out to be roughly, we think, two or three business units. You should think of us as doing this every quarter. You should expect a drumbeat from us of every quarter doing another unit.
Any other questions from the audience? Tom, going back to margins, why is 35% the right margin if you have this big opportunity? You could be doing 30% margins, which would still be amazing. Maybe drive some incremental growth.
Actually, it's an interesting question for the industry, not just for OpenText. I think many of you in the investment community with AI first should be demanding that all the tech companies do more. I think Rule of 40 should be replaced with, you know, I mean, the management team at OpenText won't be happy. I think all of you should be demanding that with AI and increased efficiency, you should see something better than the Rule of 40. What will it end up being in terms of the growth and the EBITDA? I can tell you with AI, we're reducing everyone from developers to RFP responses to legal. A lot of this is being, and you're going to see that filter its way through society, whether you're Marriott Hotels or you're Bank of Montreal. You're going to see that. Where it ends, I don't know. It'll be more.
I just, I don't have a good enough feel yet. Rule of 40 has been around for a long time. Will we go to a Rule of 50? Probably. How soon will we get there? I don't know.
Almost out of time. Maybe one last one I'll squeeze in is, does it still make sense for you to own your own data centers? Just, you know, CapEx intensive, a lot of other software companies are in.
That's a great question. You know, if you had asked me that three years ago, I would have said, no, we're mothballing or reducing our data center footprint. All you have to do is look at what New York Times is suing OpenAI and Microsoft. They're suing them because I suppose they're worried that tomorrow morning's New York Times could be written by the bot better than their editorial team. That's not a slight on the New York Times. That's the reality of what's going on with Agentic AI. These bots that we're deploying for customer support are better than our 30-year best employee. This is why this will be adopted at such a fast rate.
If you then think about that and you think about Marriott, et cetera, if you're a CIO and you're talking to your CEO, do you necessarily want to have the crown jewels that go to the rack rate or the occupancy? Do you want that up in a hyperscaler? I think we're starting to hear, and one of the reasons why you're seeing our licensed revenue, which we thought was going to go to zero, may not. I think we may be moving into a world of hybrid. That's why I had said before Aviator was going to be a multi-cloud where it's on-prem all the way to hyperscaler. Stuff like your website, which is public, you don't really care if it's in a hyperscaler. Some of these things that are proprietary that are inside your firewall, you may want in a managed data center.
We're sort of watching this unfold. We have to be driven by the customer. Right now, Aviator is built on a hybrid approach where if you want to stay on-prem, if you want us to manage it in our data center, or if you want to take it to a hyperscaler, we give you the choice. We're trying to be led by the customer.
We're on time. Any parting words, Tom?
No, thanks for your interest. I hope we make this an easier story to follow.
Thanks, Tom.