My name is Seth Gilbert. I'm one of the SMID software analysts here at UBS. Today we have the honor and the privilege of chatting with Tom Jenkins. Tom was the CEO of OpenText from 1994- 2005. Now he's the Chief Strategy Officer and also the Executive Chair of the Board since 1998. So there might be no better person to take us through the discussion that we're about to have, Tom. Maybe first we'll jump into it with the CEO search. I'd be curious to know. I know we're going to have the announcement by the end of this calendar year. But as you've gone through the search, has it evolved? Have you changed kind of the type of person that you're looking for? Or has it been kind of meeting your expectations of what you were searching for?
A couple of things. One, I'm not doing the search. The search committee is doing it. We thought that all of our new board members, because we've changed over our board quite substantially, so we left it to the new group of board members. We've got the just-retired CHRO of Hewlett-Packard, et cetera. So we've got an outstanding new board: CIO of Cisco, et cetera. So we've got a wonderful search committee. So they're off doing that. I would say their thesis has been the same right from the very start. They're looking for someone that has a solutions background. We're a heavy, heavy technology company, and we need to emphasize more about solutions and things that are more compelling for customers.
Got it. That makes sense. Tom, you said nothing was off the table. So let's just jump right into it with divestitures. Companies going through changes. You've announced that you're going to divest a pretty large chunk of the business, up to $1 billion of your revenue. I'd be curious. You listed core businesses as Content, ITOM, Business Networks, and Cyber. So I'd like to dive into maybe a little bit of each. Maybe we can start with Content. Why do these businesses make up the core?
Right. Well, first off, one of the reasons why we had made the board changes and the management changes was OpenText found itself really trying to manage too many business units. At the same time, it had six of them. And for a company its size, it just proved to be too many for too small of a company. It had about $6 billion in revenue. So we sold off the mainframe business, AMC, last year. We wanted to move more quickly to sell off some of the other units. None of these business units. There's nothing wrong with them. It's just there's too many of them. So we decided, and it wasn't a very hard decision for the board. We decided that we would jump all over content that trains agentic AI.
We were blessed to have the content division as well as some of the other units we can talk about. And we started to say to ourselves, well, if we're calling on the CIO at General Motors, we probably shouldn't be selling to mom and dad at Best Buy. So consumer divisions or developer divisions that really didn't fit with that core thesis were the ones that we declared as surplus and put them up for sale. And so we're in the middle. We've sold one unit so far. We're trying to sell one a quarter because that's the reasonable number to sell. We think we'll do that over the next three quarters. And you have to do that because as you divest a business, it takes a lot of accounting work and a lot of separation work.
And we didn't want to harm the core business while we were doing it. Now, the core content business, well, that was a no-brainer. As someone said to us earlier today, usually when you hear this story about somebody cutting their company down to the core, it's usually the core is slow growing and it's highly profitable. We're actually cutting down to a core, which is our fastest growing and our most profitable. So it was a pretty easy thing. We hope when all this is done that we will have reduced about 15%-20% of the company. So we'll go from what is now $5 and change billion to $4 and change billion.
At that point, the core of the core, as they like to say, the Content Cloud business, which is about a third of the overall content business, it's growing at last quarter 20% plus growth. So it's good, solid double-digit growth. Makes all kinds of sense. So you get a better sense of us when we're down to the core. We're really a cloud company growing at double- digit. That's our plan. We'll see how we do in the next year.
Yeah. That makes sense. Maybe just one follow-up on divestitures. A question we get from investors sometimes is we appreciate all the color and the clarity that you and the company have been providing. But why aren't you able to move faster? Does it create too much of a distraction to do, say, two a quarter? Or maybe it just takes a little bit too much?
Well, there's selling and then there's divesting. And let's separate the two. There's enough buyers to sell all the units this quarter.
Wow.
But the problem is you can't divest them all because you have to, you know, it's like unscrambling an egg. You've got to be careful that you have service level agreements with each other. So it's actually not a market demand issue. It's really an operational issue. So you should be very careful. Over the years, I've sat on various boards. And you have to be very careful when you're discontinuing a business that has been together for so long. So it's more of an operational caution to do it over time like that.
Got it. That makes sense. During the most recent 1Q September earnings, you provided a new revenue breakout, which was very helpful by business unit. And it was great for investors to see because it's something we've been asking for, I think, for a while now. You have Content, Business Networks, Enterprise, Cyber. These were all growing above the company average. But when we take a look at ITOM's growth, right now it doesn't appear to be growing above the company average. So I guess the question is, are there parts of ITOM that maybe are growing faster than other parts? Or I guess what I'm hinting at is, why is ITOM part of the core?
I brought a prop along with me to understand ITOM. Go grab this book. This is old school. It's in paper. We provided this at our user conference two weeks ago in Nashville. But you can go online and download this. This answers that question. ITOM, IT Operations Management, in case you don't know, and other things like Business Networks, the Content that we talk about is Enterprise Content Management. These are all acronyms from a previous era of enterprise software and all very important areas where we were solving in the enterprise very specific problems. IT Operations Management was generally led in the early days by BMC and then later ServiceNow, Computer Associates. I watched the whole thing. I sat on the board of BMC, so I know it well. These are sort of now artifacts of a previous era.
What matters, and if you read this book and others like it, it's all about the data. It's all about the content that trains an agentic AI. Because as you may know, MIT came out with their study on the performance of early chatbots. And it's been colossally disappointing. Only 5% of the agentic bots that have been created inside the enterprise have succeeded their project goals. The two reasons why MIT cited, number one, they didn't have the right content. Remember, most of the public chatbots are trained on Reddit, Wiki, public information. If you have to go inside the firewall and you're making a bid and you're a hotel responding to an RFP and you want a bot to do that, you need to know things like occupancy and rack rates and what you did for that convention two years ago in Vegas.
Those are all things that are not in Reddit or in a public domain. You have to train with the information inside the firewall, so that brings us back to ITOM. ITOM is a critical component of three kinds of content. Human content is generally regarded as ECM, Enterprise Content, which is the origin of OpenText. The second content is actually transactional content, and that's generally content that originally came from EDI, and so that was electronic data interchange, so all of that is what's in business networks, and the third kind of content is machine-generated, and machine-generated is part of ITOM. You have to have all three types of content if you're going to train agentic AI, whether you're in an automotive company or a pharmaceutical company. That's why those are core.
Now, in ITOM, what we've done is the business that we inherited from Micro Focus was an on-prem offering up against ServiceNow, et cetera. We've now intr`oduced full cloud. So all of the things that we just talked about, Business Networks, the old names, because I think we will soon not use those old names. We will talk about machine-generated content, human-generated content, and transactional. All of those at OpenText are now in the cloud.
Got it. We can stay on cloud.
Ok.
A bright spot of the most recent earnings was cloud growth accelerating from 2% to 6%. Some of the rationale was that customers were, I guess, pulling OpenText as opposed to OpenText pushing customers to go to the cloud. You guys were being pulled along, and customers were saying, we'd rather go to the cloud than stay on-prem. The question that we've been getting from investors is, why now? Has AI been in the market maybe a little bit longer, and so a lot of your customers are seeing some of the benefits, hearing, reading about it, and wanting to go to the cloud? Or maybe there's a different reason that in 1Q customers decided more than in the past, hey, OpenText, we'd like to move to the cloud.
Well, if you think about what we just talked about with the MIT study, there's no question that customers want to have all that content surface so that they can train an LLM. So that's no question that's a forcing function. Prior to that, our customers, SAP, Oracle, Microsoft, et cetera, all the enterprise customers, they were all moving to the cloud. But that was generally motivated to save money. And so there was a very good ROI rationale for doing that. We tended to leave it to customers to come to us and decide. That resulted in virtually very little of our maintenance space actually switching over to cloud. And that the cloud revenue growth that you're referring to were really new name account growth. And it really wasn't part of a program.
Now, Steve Rai, our CFO, signaled at the most recent quarter that we're going to probably start changing our revenue mix. Now, that confused people a little bit. They said, well, what does that mean? Are you off your plan? And we're having our plans fine. But what we're doing, and at our analyst day, people can go and see the slides that Steve created, you'll start to see maintenance start to come down and be replaced by RPO. And he was starting to run everybody through the accounting. It's the accounting that Salesforce and SAP and others do. So there's nothing magic there. But people have never really seen it at scale from OpenText. And what Steve was trying to indicate to people, start to prepare, you're going to see even more cloud growth.
And I think the objective will be you will see us have majority of revenues from cloud. And you will also see our maintenance conversion. And quite frankly, we're going to follow the playbook that other enterprise software vendors have used, SAP, et cetera, where you'll see a dollar of maintenance replaced by multiple dollars of cloud. And you'll see a lot of installed-base marketing from OpenText. So there's a lot of growth there. And it will be a very intentional strategy. And there'll be more from Steve in the quarters to come as he starts to get a better idea about margins and how all that works out.
It's something that investors are asking us about now. It sounds like you're still formulating how to communicate to the street. But maybe one more question about that would be total revenue in 1Q was a bright spot. Came in above our estimates, above the street estimates. But the guidance was a little bit light. Does this transition that we're going through right now on the license to cloud side, does this have anything to do with maybe the 2Q coming in a bit light? Or maybe the street was just modeling a little bit too heavy to begin with?
I think it's revenue mix. Because what happens is when you swap $1 of license, it actually is only $0.40 in that year. But you see the rest of it. So I'll do basic math. I know Steve will kill me if I go into too much detail. But say you had $1 of license and you swapped it out for $2 of cloud. The way it works out is it works out over many years. Your minimum contract value is three or four years. So you do get a bit of a J Curve effect. You end up with way more money and way more profit. But in that short term of 90 days or 180 days, you'll see your revenue mix change. He was trying to signal that.
In the maintenance example, you don't have a J Curve because the $1 of maintenance draws down and is immediately replaced by what's called CRPO, which is your current remaining performance obligation, which will be even more. So license will shift down lower, but RPO will go up much larger than the shift down. So it does become a timing issue. And I think that probably either caught people by surprise or confused them. But that's why he did the charts on analyst day so that everybody can, and as you said before, we're trying to be as transparent as we can be so that you can track all of it and see, Ok, I had a $1 come down here, but now I see $2 up on a different part of the balance sheet. So there is a bit of, let's say, educational communications we have to do.
But make no mistake, we are going to the cloud. And we're going to go with much faster growth because that's what all our peers have done.
Right. Maybe you could talk a little bit about some of the deal sizes when customers move to the cloud. We could start with current customers. So are you finding that I understand there's a difference between the revenue recognition? I think the audience and investors is probably well understood as well, seeing other models. But are the deal sizes starting to increase? Maybe they're adding on AI. Maybe they're buying more products from you. It's a chance for your sales folks to be able to educate the customers on what else OpenText has to offer.
So this is an excellent question that we are late to the party. So we can, and in fact, have learned from many of our SAP's install base and our install base are almost identical. And SAP's been a great customer of OpenText and partner and vice versa for decades. So we're learning from them and from other partners at Microsoft and Oracle, of course. I would say we have to crawl before we run. So we will model, I think, safely that we will do $2 for every dollar. I think we're quite comfortable that we can do that. Shannon Bell, our CIO, showed, quite frankly, more than $2 of value off the dollar maintenance in her keynote at the user conference. But for now, I think what you'll see us say is as the dollar maintenance comes off, that we will earn our way to $2.
Other vendors in our peer group that have been at this longer than us are now doing $4 and $5 because they've done additional applications, et cetera. I think we have to crawl first before we can run. But as we move into that cloud platform, we have Aviator. We can start to add on those applications, specific agentic bots, et cetera. But we're early days. So I think we're comfortable to say 2: 1. But obviously, our peer group is doing much higher than that. So we hope to, it's a process. It's a journey that we're on. So we'll have to learn how to do that. The good news is it's all the installed base. So it's not going anywhere.
I think that's helpful. It's a helpful fact. Maybe the last one on cloud. As you've always had a cloud offering, as you have current customers who are moving from on-prem to the cloud, does that open you up for new customers coming in and the ability for them to recognize that maybe your cloud is a little bit more advanced than it was yesterday? Or maybe that's a little bit too much of a stretch.
I would say that it took, with the Micro Focus acquisition, it took two years to bring what was a large library. I mean, it is Hewlett-Packard's original software library. It's an enormous, wonderful library. But it took two years to bring that to the cloud. And today, at our user conference two weeks ago, we now have all our major products on the cloud. So everyone has a choice. The question then becomes features. What is a true feature parity? So in some regards, we still have to achieve feature parity. But as of today, though, all the main product lines and the Aviator data platform that we announced at the conference, they're all sitting there in the cloud now. So customers have that full choice. We already have 10,000 of the Fortune 10,000. So it's not like we're going to get new name accounts.
But I think people will certainly look at this, especially from the following point of view. When you have a unified data platform, you can start to take all those things we were talking about before, whether it's ERP, ITOM, CRM, et cetera, and start to have a unified view of the content so that you can train a chat. An agentic AI does not care that the data you're educating it with came from an old ERP system or CRM system, et cetera. What the challenge for all enterprise software companies is that that's the old world. And we have to present data that an LLM can digest. But it's even more complicated than that. And that's why I wrote the book, because you can't just do that in the public domain.
One of the reasons why OpenText exists today in enterprise content management and enterprise information management is that most of the stuff's behind the firewall. And you can't just make that available. And if you educate an AI and you allow the public to query that AI, you've just provided them all the information that was inside your firewall. You have to have an architecture where you can govern the data and govern the AI, because an AI cannot unlearn something. If you're going to unlearn an AI, you have to start all over again and go through all the power consumption, go through all the training, and literally start from four walls and a telephone, as they would say. You have to start all over again. So a key thing here is the ability to handle public content, private content, and then partner content.
And the hardest one of all is the partner content, because it's sort of in a demilitarized zone, a DMZ. And it took us a decade to figure out the permissions governance model of, say, you're a General Motors and you've got to work with your dealer network. Well, you have to share proprietary information. But it can't go to the public. Or say you're a tier one, tier two supplier to Pfizer. Well, all that information, those clinical trials, all that stuff, that's got to stay private, even though you have to work together to do an FDA submission. So this is a very complicated world. You've got to be very careful about how you go about and orchestrate and handle all that. But the reality is, for corporations, it's an enormous productivity gain. But you still have to work with those different types of information.
Got it. There's something I wanted to follow up on, too. You talked a little bit about security, or you're hinting at security. In the past, about a year ago at OpenText World in 2024, I believe, I guess Mark had mentioned, CEO had mentioned how security is a layer to wrap around everything and about how every customer could be a security customer. That journey hasn't exactly, in the past 12 months, I don't think it's exactly played out to be a security customer. But with sort of the reinvigoration that's here right now for OpenText, maybe you could talk a little about cybersecurity and why cybersecurity is an important element of its core.
He was absolutely right. And I learned something, actually, because having sat many years as a CEO and on boards and what have you and chairs of different committees, I was perplexed by this because I thought, well, everyone already has security products. What I learned from CIOs is you can't have enough security products. In fact, what was happening, I thought that when we would put the security wrapper around content, around business networks, et cetera, that it meant that somewhere else that there would be security removed. No. What happens with security? You do layers of security. And if someone tells you you've given them another layer, another wrapper of security around a particular content or application, they're actually quite happy to have multiple layers of security. That's actually what's happening. And so now there's a difference in OpenText.
We have a business unit that does security for PCs at Best Buy and consumer division. And then there's security at the enterprise level. And so I'm speaking now at the enterprise level.
Got it. You made some comments, some interesting comments about OpenText being the most open data platform. I think it got some airtime, but maybe not enough. I'd love to turn it to you to tell investors why that's important and is that a competitive advantage for you.
So in OpenText history, we've always been known as Switzerland. And we did that on purpose because many years ago when we were building search engines, people wanted to leave their content in its native form. So back then, if you were in Lotus 1-2-3 or if you were in WordPerfect or you were in Word, and then later it became Google Docs, you wanted it to stay in your native format. So what we did is we said, look, we will crawl all of this data. We'll go into your native format. So today, OpenText has over 1,500 connectors to everything you could possibly think of that we built up over the last 35 years. And so you can go into Lotus 1-2-3 and you can have a viewer come up and show you VisiCalc or whatever it is.
That was so important to building an enterprise content management system because you cannot go to a corporation which has been doing FAA filings or FDA submissions, et cetera, and say, oh, good news. We've come up with a new software program. And we'd like you to get rid of everything that works just fine. And of course, the organization's going to say, no, we're not doing that. And so it's important that you have this sort of Swiss Army knife approach where you allow all those pieces. Now, fast forward now to training an agentic AI.
That's why Aviator is so important as a data platform, because we've taken that philosophy and said the same thing, except not just the content, but also the large language models, because we're not presuming that you're going to just use Anthropic or you're just going to use Microsoft or just use Gemini or whatever. We're assuming that organizations will pick. And they will pick different models. So we've made Aviator so it's multi-model. And we've also made it so it's multi-application. If you look at the MIT study, the other thing they remarked on besides the behind-the-firewall content was the bot would make the right decision, but it wasn't connected to anything.
So here it was making the decision based on the information, but it didn't have the workflow that allowed it to go into the ERP system or in the CRM system and actually cause an effect. So it was doing all the work to make the right decision, but nothing would come out the other end. So making all those connectors is another part of being a Swiss Army knife. So Aviator was made so that it could work with any large language model. Keep in mind, there's 400 of them, viable LLMs right now. This is crazy. A year from now, we're not going to have 400 LLMs. But we will have LLMs that are tuned for an industry, for a nomenclature, for a particular analytics engine problem. And as the customers choose that, we will interface to those, just like we did with customers choosing content.
So the idea there was to build a Swiss Army knife, which really left the decisions to the customer.
Got it. Another one from OpenText World. Last month, it was packed with announcements. You had the AI data platform, Databricks partnership, Aviator Studio, Aviator AI services, probably a few others that I didn't mention. This might be asking you to pick a favorite child. But can you tell us maybe one or two or a few that you're most excited about?
Studio, I would say. So the book that I mentioned here, this is really an architecture book. So as an engineer, it was a lot of fun to work with some of the execs at OpenText to write this, which was to sort of explain away some of the myths about training AI and how you would go about it. Studio, Aviator Studio led us to a whole bunch of other concepts that people are just starting to grapple with. And you're going to see a companion book come from us, because we started talking to CIOs about who's training the bots. Are we doing it? Are you doing it? Or is the system integrator doing it? No one knew. So we started a project with many of our customers to say, OK, if you're an accounts payable clerk, do you really need to design that?
Do we really need a systems integrator? Or does that just come as part of Studio? And so we've started to have those discussions. So if you're a pharmaceutical company or a health care provider, when you get deep down into claims processing, which goes to the heart of their business, they want to train that bot. If you get to something at an industrial level within an automotive manufacturer, then someone like a PwC or a CGI, et cetera, they could train that bot because it's an industry subject matter expertise. So you're going to see, just as we talked just now about content and the architecture, you're going to see us sized with the genome of a corporation in terms of its bots and how many agents.
I'm betting right now, from two weeks ago, I think you're going to see us come out with a book that describes 1,000 agents in a genome. It'll be for a 100,000-person corporation, because the cutdown will be by role. It'll probably be something like 100 people to one agent, something like that. You're also going to see probably somewhere between 50 and 100 orchestrators that connect all those agents together. So that's a preview on the next book we're working on. So Studio is my favorite.
Got it. I think that's a perfect time to close. We're out of time. Thank you so much for joining us, Tom.
OK. Thank you.
Some books here.