For joining day two of the tech conference. We have PTC here, and excited to have a product and practitioner focused discussion. I think a lot of exciting things going on in the product development, AI impacting a lot of different industries that PTC serves. So excited to welcome Kevin Wrenn, who is the EVP of Products, and we also have Michael Maguire, newly appointed IR Chief from PTC. Gentlemen, thanks for making it down to our conference. Kevin, maybe just for folks in the audience that are less familiar with you, give us an overview of your background. What are the areas that you're focused on at PTC?
Sure. I have been at PTC for a long time, over 30 years. I have spent the first half of my career, let us say, implementing our products. I was the Head of our services organization for a decade, and then I have spent the second half of my career in products. I was the General Manager of PLM, and now I am EVP of Products. These days, I spend most of my time customer-facing, sponsoring our most important clients.
Great. I want to talk a lot about the announcements this week or, sorry, this year. You recently hosted a big product launch event. Maybe just give an overview grounding for the audience here of the PTC product offerings. You have talked about intelligent product life cycle in the past, so maybe some more details on that to start.
Sure. PTC's customers, we service customers across a bunch of different verticals. Think industrial, FA&D, electronics and high tech, medical technology, automotive. They are our main customers, and they share a certain set of characteristics. That is, our customers make highly complicated electromechanical equipment that is difficult to engineer, it is hard to manufacture, and we say have a long and interesting service life. When we talk about the intelligent product life cycle, let us say the basics of it is that these companies, let us talk about, let us say, a rooftop air conditioner manufacturer, Trane or Carrier or someone like that. Their engineers are in charge of engineering next generation products, and that is their biggest focus. But in the end, engineers produce data for other people. Engineers produce data for manufacturing organization, for the service organization, for the marketing organization, for customer success, et c.
Our strategy around the intelligent product life cycle is that for companies that share those characteristics, it works the same way. All of that data originates in engineering, and when changes happen, generally it emanates from engineering. Let me give you just an example that I like to use with customers to explain to you what it is. If you think about a mechanical part, any kind of mechanical part, it starts as a set of requirements in, let's say, Codebeamer or a requirement system. Then an engineer will make a CAD model. They'll put PMI information on it, which is, let's say, manufacturing information. That part gets to be part of an engineering bill of materials.
Then the engineering bill of materials manufacturing engineer will turn it into a manufacturing bill of materials, again in Windchill, and a set of service instructions, or excuse me, manufacturing instructions, and that data will travel. One place it'll go is to ERP. When it gets to ERP, it turns into something different. Either it's called a material master or it's called master data. From there, supply chain people make derivative works, order parts, pick suppliers, et c. Likewise, they send the same information to the manufacturing floor, to a manufacturing execution system to build the product. That creates an as-built bill of materials that you could send to a place like ServiceMax or Orbit to be an as-maintained bill of materials.
This is just a simple example of how data, let's say, travels from engineering to all of these different places inside of the industrial enterprise for the purposes of speed and quality.
Yeah.
That's one side of the intelligent product life cycle. Let's say, forward propagation of data. The other thing is, okay, how do we react when there's a change? Let's imagine that the customer did all of that perfectly, then someone decides they want to change a supplier. The first thing they have to do is they have to say, "Okay, well, what's the impact?" The first questions they'll ask themselves is, "Okay, where is everything? How many parts do we have in factories around the world? How many parts do we have in service stocking locations around the world? How many requirements have been built specifically for this supplier?
How many designs in process have already designed this supplier in? This is this idea of originating the product data foundation from engineering through all these other systems that house parts and BOMs across the enterprise, then the ability to react to a change. At the highest level, that is what we are embarking on at PTC.
Okay. No, that is great. I wanted to dive in a little bit more on how things have changed at PTC with AI. I think you having the privilege of over 30 years at PTC, you have seen a lot of different technology adoption cycles, but just give us a sense for how has the product organization embraced AI? What are some of the biggest increases in product velocity or productivity you have seen out of your team?
Yeah. Like everybody, we let our engineers experiment with AI. What that started as, let us say, individual productivity improvements in, let us say, requirements gathering and product managers could automatically do prototypes and things like that. Developers used it to do code generation and test case generation. The QA team started to create test harnesses and automation all with AI and all individually. Then what we discovered is it is just like a factory, you are only as good as your slowest machine. We have embarked on a more holistic approach to infusing AI end to end in the process. I would say we are in stage two.
We are beyond individual productivity now, let us say coordinated productivity across different functions. Ultimately, we are looking to have agents coordinate across the product life cycle to accelerate things. Have we seen throughput and better quality? Yes, but we are anticipating much more as we get this much more end-to-end process in place.
Right. You recently held the PTC NEXT conference, I think it was back in June, in Chicago, and announced several new products. Maybe just give folks a quick overview of what you announced and then what has early customer feedback been. I guess, given that you haven't still a lot of these AI improvements are still to come, why do these product announcements now? Obviously you can do more later, but just give us a sense on what you announced and what customer feedback's been.
Sure. We announced two big things at PTC NEXT. One is called PTC Jetstream, the other one's called PTC Orbit. We also announced a bunch of AI.
Yeah.
You can let me know if you want me to touch on all of it, but the big ones are important. PTC Jetstream is a collaboration tool that acts as, let's say, a module of Windchill. Windchill's our PLM system, our customers evolved version, chain configuration control their designs inside of Windchill. We found two things that we could help drive some value on. While Windchill was the system of record, internally when engineers collaborate throughout the enterprise, there's folks that are not users of Windchill, so they have to do collaboration outside of the system. Let's say they have to collaborate with a supply chain manager. That person wouldn't normally have a license of Windchill, so the engineers would collaborate with them through meetings.
Now we have PTC Jetstream where they can actually collaborate in, let's say, rooms, spaces, private rooms that they could do all kinds of collaboration on. When they're done, they can bring that back into the system of record to, let's say, memorialize that in a configuration, while being able to take, let's say, all of the collaboration notes and things like that back into Windchill. The second thing is our customers have to collaborate with their suppliers, and right now they do it asynchronously. You can send a package of data to a supplier, they can work on it and then send it back to you and it'll find its place back in Windchill. But it's completely asynchronous. The second thing is our customers don't want their supply chain in their system of record because that's where all of their intellectual property is.
Jetstream allows you to publish a subset of information into a little bit of a private collaboration room, just like you would do internally, but now you can do it with a supplier, and when you are done, you can bring that back into Windchill. The creation of Jetstream was really more at the request of our customers than anything. They were frustrated with the asynchronous collaboration process, both internally and with the supply chain. The early feedback, it is in beta right now, has been very good. We are excited. We are going to launch it here in October.
Great.
The second one is called PTC Orbit. For forever, or for at least as long as I have been around, a lot of our customers have wanted to manage an as-maintained bill of materials. Meaning, while the product is in the customer's hands, you understand what the change in the bill of materials is while it is in use, for two reasons. One is so you can improve service. When technicians go there, they can know what they are looking at, know what they are going to work on, they know the condition of it, so to improve service. Second thing in service, so they could run campaigns throughout the install base. Then on the other side, we will use the rooftop air conditioner example again.
When engineers want to engineer the next generation of product, they sure would like to understand the repair history on the fleet or the different performance characteristics of the fleet in different regions, let us say. That is PTC Orbit. Actually, version two of it is coming out here next month. Again, early days, but customer feedback is good and we are optimistic.
Awesome. As we think about just the changes in the software industry in terms of packaging products, obviously PTC has primarily been a seat-based model. Frankly, a lot of software, end-user focused software has been seat models. How do you think that plays out, again, not this year, next year, maybe three to five years down the road? Is this still going to be vast majority seat-based? Is there going to be more of a usage outcome-based component as you start integrating AI that can move from a productivity enhancer to a workflow automation. How do you see that playing out?
Yeah, I think the answer is yes and yes. For the products that you all know us best for, your Creo, your Windchill, your Onshape, your Codebeamer, those are going to be predominantly still seat-based pricing. However, when we think about the rollout of AI in various different levels, that's obviously going to add a consumption-based component to that. We're going to evaluate the types of AI that we're deploying out there. There's lots of different pieces. I think inevitably there's going to be some level of AI that will be expected by our customers to be baked into a seat-based price.
Yeah.
That's just inevitable, not just for PTC, but I think across the board in much of software. But as the complexity and the ability of that AI technology that we deploy scales over time and does more complex things, there will be a consumption-based component baked on top of that. So initially you would go and you would purchase a certain number of tokens associated with whatever AI SKU that you were buying, and then once you worked your way through those tokens within a given period of time, there'd be almost an overage consumption-based model on top of that where you would be required to purchase your own tokens. From an outcome-based perspective, we're not necessarily there yet.
I think it's an interesting idea, but we've had conversations about how you have to be very specific and succinct about what the actual outcome is you're driving, and you have to be in lockstep with your customer on what that outcome is, so there's no issues when you go to charge them for it. So we're not necessarily there yet. I don't know if we will be there in the next three to five years, but it's something that we're exploring and considering. Then the last piece would be in a world where agents obviously start to work with some of the workflow, a little bit more external agents, we do anticipate charging in a seat-based way, in the same way we would charge the human. Then on top of that, we have monetization tools with API connectivity as they ping into our system.
Overall, that's broadly what we think about from a pricing perspective.
Okay. Have you started to introduce some of the usage-based pricing today, or is that maybe something you expect over the next year or two?
Yeah, we have. We have products like, our ServiceMax AI product would be one of those examples, where there is a certain level of consumption that obviously is baked into a seat-based SKU for that ServiceMax AI SKU, but then beyond that, there would be consumption-based revenue that would be generated.
That model will be consistent on, let's say, embedded AI inside each of the applications.
Yeah. Got it. I guess just as we think about some of the product level changes or even the way that customers are deploying, consuming PTC software, how much of that needing to move to the cloud is a prerequisite for AI? A lot of customers run PTC on desktops, their own servers, and a lot of these companies are complex, regulated industries. Do those customers need to migrate their data to the cloud, or their CAD and PLM environments to the cloud to take advantage of AI? Or how do you think about that?
No. As you said, the vast majority of our install base is on premises.
Yeah.
When we access large language models, it's hybrid. Their data will stay on their tenant and access, let's say, the large language models through the cloud.
Okay.
I would say all of our LLM works that way. If you're a Windchill user, you can be either on premises or in the cloud, and you can leverage AI. You can also get it through our subscription of Azure. We can also make it go through their own subscription of Azure, but it could be hybrid.
Yeah. I guess going back to maybe some of the internal stuff you're doing with AI, what are some of the big tools that you're using within the product and engineering organization, and are there specific processes that maybe you could talk about that have seen the biggest ROI or productivity gains?
Yeah. I guess the measurements have been vague, to be honest, so far. You see very interesting thing happening. At PTC, the way things work from a product manager, a product manager is market facing, and then we have something called the technical product manager, converts that into things that R&D folks can work on. Now, all of a sudden, because a product manager can use Claude to do a prototype, a lot of those things streamline the communication with the developers and starts to make us think, "Oh, okay, can we now collapse that role into a single role and take a bunch of time out of the process?" It seems like that's happening right now in pockets, and I talked about that before as trying to, let's say, institutionalize that or make it more of a durable process. That's happening.
A lot is happening on code generation and code, let's say, quality assurance. The other thing that we're seeing a huge impact is response to customer issues. We get into really complicated customer issues and really sophisticated IT environments. We now can solve something in technical support that will take hours vs what sometimes would take a month to troubleshoot. We're seeing improvements all along the life cycle, and right now we're seeing everything through these keyholes. What we're trying to do is, let's say, make it uniform across the process to get, let's say, rising tide to lift all boats.
Right. I wanted to ask you, as we've just seen a lot of headlines over the last week with some of the frontier models, Astra release, some other kind of AI-focused CAD models coming out on the market. How would you characterize what's been released, what is coming, and is there an opportunity to partner with them, or how would you just sort of distinguish what the frontier is building that some investors are viewing encroaching on your space, vs the reality of when you're speaking to customers?
Sure. From our perspective, we're excited about the developments of large language models and CAD. Fundamentally, if you think about just the mathematics of CAD, it's deterministic and it's complicated to, let's say, resolve mathematics on edges and surfaces and things like that. Those are all, let's say, mathematics. We don't think that training on large amounts of data of CAD models is going to work. What we do think is that what LLMs are really good at is generating code. What we can do is we can use code to generate CAD models. So rather than text to CAD, and I'll tell you some other reasons why I think that won't work. We're saying text to code to CAD.
If you can go to NEXT and you can see a presentation by the Onshape folks about FeatureScript and doing designs in Onshape with that same motion, code to text to CAD. So the code creates, let's say, API calls that uses normal end-user motions to create a CAD model, which will let you automatically produce a CAD model, but it also will produce a CAD model that an end user can adjust. We also think that there's a lot more to CAD than just getting a shape as quickly as possible. You also have to consider, well, we're building it for our own factories. We're building it because we want to incorporate these suppliers. So for us, we think the future of text to CAD isn't text to CAD. We think it's text to code to CAD.
I think for the CAD industry, not just PTC, it creates an opportunity because why would anybody want to try to recreate the complicated mathematics that the leaders in CAD have figured out over the last 40 years, rather than just figuring out how to leverage it and drive those models? Furthermore, on our internal research, doing engineering isn't just CAD. Doing engineering is engineering. You need to do all other kinds of things like understand the environment that it is going to be in, understand the reliability requirements, manufacturability, et c. We think large language models can help with that as well. Our research is proving that out. What it also does is produce a lot of other data that you will need for compliance, regulatory, explainability reasons of why you came up with the design.
We think that text to code to CAD also creates a data management opportunity for things like ALM and PLM.
That is an interesting way of framing it. As we think about some of the core products, maybe stepping back from the AI conversation for a moment. As we look at your CAD and PLM franchises, they have obviously been dominant franchises out there in the tenure that you have been at PTC. I think you guys have, at least based on my remembrance over the last few years, have called out more competitive wins and replacements, and I know some of your competitors in that space have struggled, maybe distracted with other M&A. Can you just frame for us how you see the market share competitive dynamics within CAD and PLM, and does a CAD or PLM migration, does that happen a little faster now because of some of these AI tools that can assist with that migration?
First, to answer the last part of it. We have not seen AI magically help convert CAD from CATIA to Creo or vice versa.
Yeah.
We have not seen that, although if that did happen, it would fundamentally change the market, I think. But from our perspective, we have two CAD systems. We have Creo, and we have Onshape, and our goal there is to take a disproportionate of shifting seats there. Folks who are coming off CATIA or coming off of SolidWorks, which seem to be the most vulnerable. So to win a disproportionate amount of those, of course, Onshape is a displacement business 100%, practically. So that's CAD, and we're after that. And we have some success both with CAD and with Creo there. On the PLM side, displacements do happen.
And it's when companies are reevaluating their PLM choice, either by they implemented their first PLM system in the early 2000s, and they realized, okay, the system itself is antiquated, or their implementation is flawed, and they want to build this digital foundation, and they decide, "Okay, we're going to go out for bid." We won a deal at a large medical device manufacturer a couple of quarters ago that went through that. And we're seeing the rumblings of this happening as people evaluate their digital transformation. And some of it actually comes on the heels of the SAP S/4HANA migration.
When they're looking at that and looking at consolidating, let's say, all of their ERPs down from multiple ERPs to one, they're starting to say, "Okay, well, another big part of our data estate here is PLM and ALM, and we'd like to consolidate that down to one." So we're seeing some real opportunities there. That's from the displacement standpoint. For PLM anyway, the biggest part of the growth opportunity is expansion, actually.
Yeah. And you touched on ALM a little bit, but would love to ask you about Codebeamer, which I think has been one of the stronger growth areas and, particularly in the automotive space, and it's clearly, get in a modern vehicle today, the complexity of the electronics, the software, is orders of magnitude larger than just a few years ago. So how would you characterize where we are in the Codebeamer adoption cycle? What does your share look like internationally vs the U.S.?
The biggest share we have in Codebeamer is in Europe.
Yeah.
Mainly because that is where they were founded. We have good presence in automotive because that is what they focused on. We are at BMW and VW and TMC and now at Mazda. We are also in tier one suppliers like Schaeffler and ZF, etc . In automotive, we are going to continue to go there. One of the interesting things is, we are winning with Codebeamer in places that PTC has never been before. Renault Cars, for example, is a Codebeamer. Mazda is a Codebeamer customer, not really a customer of any other. There is opportunity to cross-sell the rest of PTC in there. Automotive is still an expansion opportunity for us. The automotive supply chain is still an expansion opportunity for us. The next industries that we seem to have some traction are in safety and compliance-critical industries because of the strong traceability in Codebeamer.
Think medical device. FA&D, anything that rolls. Those are good opportunities for us. Of course, there is a big opportunity here in the U.S. with cross-sell into the Windchill base of Codebeamer. I would say the first tranche of Codebeamer was to go into the customers that Codebeamer had landed and expanded. That is what we did at VW. The second one was to try to win new automotive customers in places where we can, like in Japan. Now the big thing is cross-sell into the Windchill base with this idea of, we call it integrated product engineering. Connected requirements management with PLM.
I see. Okay. Then maybe touching on ServiceMax, I think that business has faced some challenges with some elevated churn. Maybe seems to be turning the corner a little bit. What were some of the challenges you have seen with that business and what is the biggest opportunities going forward?
Yeah, I think I will start with the challenges that we had before, and maybe you can talk about the opportunities. I think some of the challenges that we saw in that business was, as we have discussed before, we did have elevated churn over a two-year period as some of the contracts that came up for renewal ultimately needed to be downsized in terms of seat counts and things like that. We have been going through this go-to-market transformation for the last two years with our customers on really meeting the customers where they are and really trying to listen and understand what their needs are across our product portfolio. Part of that did, in turn, come from some churn upon renewal. We do feel like we are through the bulk of that and the vast majority of that kind of elevated churn.
We do anticipate ServiceMax to still be dilutive to growth, but we are excited about some of the opportunities with ServiceMax AI and others. Then you can talk about those opportunities there.
Yeah. I think ServiceMax AI and the install base is a really good opportunity. The team is really accelerating roadmap there. Then, we just have to figure out the cross-sell into the Windchill base. Really. That is the thing that we have not yet quite figured out, and this idea of Orbit and as maintained bill of materials is an attractive thing to our customers.
Yeah. Okay. That makes sense. As you think about just the dynamics, maybe this is a question for Michael, but on the way that you are approaching customers, from a contracting perspective, right? There were some changes with this move to CPP and new sales leadership, obviously new leadership of the company, CEO, CFO, over the last few years. I guess there is maybe two questions. One, can you just unpack the financial implications of that move to CPP, how investors should be thinking about that layering into growth? Then maybe the second, more strategic question for Kevin, is just the way that you are engaging, the go-to-market team is engaging, with PTC customers, how is that enabling you to get in and have more conversations and basically showcase all the things that you are doing on the product side?
Yeah. So the changes that we made from a pricing perspective was we had a process called NTE 8, Not- to-E xceed 8%, which was essentially a negotiation with the client upon renewal on what the pricing uplift could potentially be at renewal. We typically drove roughly 1%-2% annual pricing uplift on renewal in those negotiation processes. We did an evaluation of our overall pricing out there in the market, and we felt comfortable with shifting to a process that we call CPP, customer price protection, which is a baked-in realized pricing uplift annually of 3%-4% into our contracts. The majority of our contracts are three years in nature. So this shift really went in full force in Q2 of 2026. We started doing it with some clients as early as last year, but in full force, Q2 of 2026.
It will take three years, essentially, for that to fully flow through the entirety of our client base. But so far, we have seen pretty reasonable success in terms of being able to get customers comfortable with this type of pricing model moving forward. Then maybe you can talk about some of the motions that we have seen overall, as you said.
Yeah. The big thing that I get involved in is expansion, and expansion to do this data foundation we are talking about. I often get involved in value conversations with customers and with our go-to-market team. Customers always put value into a few categories when they are talking to us. One is engineering productivity, one is cost of goods sold, and the other one is usually cost of poor quality. Oftentimes, we are talking to the engineering folks, and they are trying to center the value of a PLM system on engineering productivity or amount of engineering hours. The conversation we will have with them, we will say, "Okay, well, how much do you spend on product development overall as a percentage of revenue?" Our whole install base will fall in a range, somewhere between 2% and 7%.
Then we say, "Okay, how much do you spend in cost of goods sold?" The range there is usually, like, 45%-65%. The whole idea is, let's focus your value case on getting after a metric that you can have a huge impact in the company vs trying to say, "Okay, great, our engineering productivity is improved by 10%." Not a soul would notice. But if you can make an impact on cost of goods sold or cost of poor quality, it makes a huge impact. So those kinds of conversations are driving expansion into things outside of engineering related to PLM and ALM and other things.
Okay, great. Well, it looks like we're right at time. This was an awesome discussion. Kevin, thanks so much for joining us, Michael as well, and appreciate everyone attending the session and we'll wrap it up there. Thank you very much.
Thank you.