← 返回

2026.09.24 | TB | Software | Software AG Customer Contracts

原标题：Software AG – How Are Customers Navigating Contracts & Mainframe Stability?

#software

Third Bridge · 2026-09-24

Analyst-led • 9 Sep 2026

Companies: IBM US

Alessandro Coppola (AC) Chief Information Officer, Retail & Banking Technology at ING Bank NV

Philip Atkinson (PA) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

PA [00:00:03]: Welcome to Third Bridge's Interview entitled Software AG – How Are Customers Navigating Contracts & Mainframe Stability? I'm Philip Atkinson, and I'm here with Alessandro Coppola, Chief Information Officer, Retail and Banking Technology at ING Bank. Welcome to the Interview, Alessandro.

Before we get started, could you please state either I agree or I disagree to the following statement: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

AC [00:00:47]: I agree.

PA [00:00:49]: Could you share an introduction to your background and your time interacting with Software AG?

AC [00:01:00]: Yes. I've been working in technology for financial services in the last 16 years in different capacity, as Chief Architect, as Chief Technology Officer and currently as Chief Information Officer. In my career, I've been working with Software AG across their product portfolio, both in the area of let me say, method for mainframe development Adabas & Natural, but also with their integration products like webMethods and also the more enterprise product for the architecture and process teams, it comes to Alfabet and to Aris.

Questions

[00:01:52] AI has caused significant price changes in software debt, factored in as a significant risk at the start of 2026, although things have been improving since February. Where do you see the AI impacts – both the risk and the opportunity – for Software AG today?

AC [00:02:20]: I think that when it comes to AI threat or opportunity for Software AG. I see actually a twofold view when it comes to their most important line of products. For example, when it comes to Adabas & Natural, it's pretty clear that generative AI represents a risk to materially reduce the cost and the complexity of application modernisation. Up until a few years ago, the idea of modernising the mainframe application or mainframe application with Adabas & Natural was considered somehow prohibitive considering the complexity of reverse engineering software code developed 20-30 years ago with scarcity within the organisation. Now with generative AI, basically, you have the possibility to get the documentation out of very obsolete code, prepare pipeline to migrate the code to another technology and get a significant help in modernising the application landscape. From this perspective, if you look at products like Adabas & Natural, AI represents clearly a risk of increasing the opportunity for clients to migrate out of Adabas & Natural. Of course, the depth of this reduction of the perimeter is a function also of which applications are currently running on Adabas & Natural because typically, this technology will be used to run mission-critical transactional system that are really at the core of the bank. I expect that in the horizon of 5-10 years, leveraging generative AI, the organisation can reduce significantly their footprint on the mainframe.

In other cases, for example, on Aris, let me say that Software AG has an opportunity to leverage the massive presence in terms of process repository as AI context. Basically, they could reach their offer on Aris that is a tremendous repository when it comes to process of an enterprise in order to augment the capability of mining, leveraging on gen AI. It depends, of course, on the strategy of the company that is today pretty static. Currently, Aris is more positioned as a system of record when it comes to process rather than a tool for modernising processes. They should be investing more in modernising the product with generative AI. When it comes instead to other products like webMethods, as you know, this has been sold to IBM and Alfabet instead was exiting the Software AG portfolio. As of today, Software AG is basically Adabas & Natural and Aris.

[00:06:06] When talking about Software AG's A&N [Adabas & Natural], you mentioned that banks will be able to significantly reduce the mainframe footprint in 5-10 years. Could you quantify that?

AC [00:06:21]: You mean in terms of percentage of the workload migration, right?

PA [00:06:26]: If we are 100% A&N today, what could that look like in a typical bank in 5-10 years?

AC [00:06:35]: Let me divide the span of opportunities in two. You can have very large users of A&N. For those, I expect that you can reach up to 60-70% of reduction because usually, when you have very large installation, you will still have really core pieces that are really the core that everybody is afraid to touch even with the proper AI assistance for a matter of risk, compliance, resilience and reputational risk or organisation that instead have more like a tight usage of A&N. Basically, they have been developing most of the mainframe leveraging on traditional Cobol. I see also possible that the migration could reach up to 100%.

PA [00:07:38]: When you're saying 60-70% reduction for those large A&N, you mean 60-70% reduction and not that 60-70% would remain A&N. Is that correct?

AC [00:07:52]: Correct, reduction.

[00:07:54] How do you think things are separated across banks, very large-scale users and smaller users? For instance, would you think most large banks are A&N users?

AC [00:08:09]: I would say that at least in the European context, I'm not aware of large banks heavily relying on A&Ns. I would say that out of A&N customer, there is a 40% of large and 60% of small customers.

PA [00:08:36]: Is that on the estate size for A&N and not the bank size?

AC [00:08:41]: Yes, yes, yes. Correct.

[00:08:53] If we've got very large customers with 60-70% potential reduction, do you think they would have more ambition to reduce their A&N footprint, or could smaller banks and users move first, given they only have a few smaller A&N use cases to address? Who do you see having the quicker start or drive to migrate?

AC [00:09:24]: There is a relapse effect in the sense that large banks have a higher compelling need because the modernisation and in one row, decommission both A&N and the mainframe consumption. In terms of business case, I believe that large bank has a broader business case to get out of A&N because they can say, "Check to get out of the mainframe and also save the significant costs from IBM." At the same time, they have the more complex journeys because they will hit first the low-hanging fruit that education return somehow more such a light and then they will progressively delve into the sports of the bank. On the other side, smaller bank has a less risky part, but also less expectation of a sizeable savings. Between the two, in terms of complexity, definitely smaller bank will have the easiest part. In terms of business case, large bank will have the best benefit.

PA [00:10:51]: You mentioned that in 5-10 years, we might see that significant reduction in footprint. Do you think we're going to see any examples of this in the coming five years, or do you think these projects will start in five years?

AC [00:11:12]: Whether something is already happening now or it's going to start? Help me understand that.

PA [00:11:19]: Is there any risk that some customers, even if they are not typical customers, could start these downsizing projects today or in the coming two years?

AC [00:11:29]: No, I think, this is currently happening because the generative AI capability have reached a significant level of maturity that could allow the transition that will allow the step. I truly believe that this is definitely starting and then I'm aware about several banks already starting this journey of modernisation. For me, the 5-10 years horizon is more a landing spot and a starting point.

[00:12:07] How long would it take for the full reduction to be seen in the A&N regarding the migration projects that you mentioned? Is it a 2-3-year project?

AC [00:12:21]: No, it's actually taking the overall 5-10 years horizon. Typically, the projects are organised by waves. Typically, you follow, on your seen approach. You start from the outer layers, which are applications that are less integrated with the rest of the system. Then you progressively carve them out, leaving the core of the transactional processing inside because that's usually the last part that you want to touch because of the intrinsic risk and the extension of testing that you need to perform. Maybe in certain cases, you may also decide that those pieces are small enough, it is rather better to keep them secured on the mainframe and on A&N without migrating them because you can eat already good business case if you decommission the outer layers of the application.

[00:13:31] Would customer spend be decreasing with every wave, with less spend going towards Software AG?

AC [00:13:44]: Basically, I'm expecting a stable-to-decreasing cost profile. It's very important also to finally prepare the business case and the contract renewal in a way that you can foresee these programmes and have them completed by the time that you need to make a renegotiation of the contract, especially because you may be in a situation where you have a committed contract for a certain duration. If you have something during the contract duration given the contract committed, you cannot actually monetise out of the transition. I see it more like to properly be planned against the contract duration.

PA [00:14:41]: How long are those contracts typically?

AC [00:14:46]: We have contracts that are in the space of 3-5 years.

[00:14:58] How would you segment a bank's typical spend between Software AG's A&N and Aris?

AC [00:15:16]: Typically, A&N represents the biggest part of the contract because typically, when you have A&N, you use them for core application processing, while when it comes to Aris, it's more specific department-by-department adoption and also the licence model that is typical by user is involving definitely a smaller number of customers. Based on my experience, a fair split would be 70/30 or 80/20 with 70% or 80% represented by A&N and the remaining part represented by Aris.

[00:16:17] What are your thoughts on mainframe technology, and do you expect banks to continue relying on it over the next 5-10 years?

AC [00:16:27]: Definitely, that's a technology that is meant to last at least for the next 10 years because usually… Of course, this is also depending on the size of the application estate that you have on the mainframe. Assuming that we are referring to mid-to-large clients. I think 5-10 years is a fair horizon and maybe for a very large bank, it's even not enough to completely get out of the mainframe.

[00:17:21] What keeps customers on the mainframe?

AC [00:17:26]: There are multiple elements. For sure, it's not something which is a purpose-driven choice because in most of the cases, clients would be happy to get out of the mainframe because mainframe is recognised as a legacy tech, and it's also leading to significant cost. Typically, the driver that create mainframe stickiness are the lack of skill, the lack of knowledge about the business logic that we have implemented on the mainframe. There are also some positive aspects because, of course, the mainframe is an icon of resilience and stability, is a benchmark when it comes to transaction throughput. It's something where business doesn't see really the value of changes. This means that it's also harder to fund migration projects because they are not considered IT for IT rather than IT for business. All of these aspects are somehow creating the stickiness on the mainframe.

PA [00:18:58]: How does AI tackle mainframe stickiness? You mentioned lack of skill and knowledge, the benefits being resilient, stability and the lack of business case. Where can AI find chinks in that armour?

AC [00:19:23]: You mean where AI can help in the mainframe exit strategy?

PA [00:19:30]: Where can AI help combat or help deal with those factors you mentioned?

AC [00:19:37]: AI is actually disrupting the mainframe application modernisation. If I look back at the experience that I had 24 months ago, the average cost for a medium mainframe application modernisation was between USD 1m and USD 2m. It was really prohibitive to tackle a project of mainframe modernisation. Now with generative AI, those sides are less than half. This means that gen AI made much more accessible, much more affordable and much more reasonable time-wise, the migration. That's the reason why it's an accelerator of the actual decommissioning of the mainframe and all the technologies that were running on the mainframe like A&N.

[00:20:57] What pushes customers away from mainframe as a technology?

AC [00:21:06]: I would say because of the risk in terms of future sustainability, because of limited skills in the shop and on the market. It's not easy to find continuity in the service. Basically, the reliability advantage that you gain from the technology, it could be offset by the underlying risk that in case something doesn't work, it's going to be extremely hard to find someone that is able to delve into the troubleshooting and get a sense of the resolution. This is definitely an element of risk. The second element that is pushing away from the mainframe is the cost because IBM is aware that mainframe is having a fixed time horizon, and they are trying to cash cow as much as possible from the mainframe usage, so that basically you can try to extract value out of the mainframe. The cost is definitely an element. I would say the risk in terms of troubleshooting and risk in terms of skill, and also, there is an intrinsic element of attractiveness of talent because for sure, talents that are playing in the space of data and AI or in the space of software development are not seeing the future or their career based on the mainframe.

PA [00:23:16]: What alternatives would banks be considering, besides mainframe?

AC [00:23:26]: Actually, there are many banks that are running out the mainframe, and they can use all banking products, which are running (? 23.38) sector like FIS, Finacle, Oracle. There are multiple alternative options to the mainframe for running the core of the bank.

[00:23:53] Could AI be a factor that enables mainframe technology by tackling the knowledge gap, where AI can be used to handle that gap and keep customers using it, or do you see AI as a push factor away from the mainframe?

AC [00:24:12]: I definitely see AI as a factor outside of the mainframe. In some cases, you can have a workaround in offloading the data out of the mainframe in order to cope with the adoption of AI also when data are into the mainframe. You do not stop yourself by offloading the data and then using them for generative AI or agentic AI use case. Most of the cases, generative AI is considered as a additional weapon to get out of the mainframe monopole that actually is definitely going on since a lot of years.

[00:25:13] Do you think mainframe migration or maintenance is a priority for banks today, or do they have different priorities, such as AI spend?

AC [00:25:35]: You mean in terms of whether AI is a priority compared to the mainframe?

PA [00:25:40]: Considering budgeting, does migration from mainframe take a priority, or do other things take priority instead?

AC [00:25:55]: No. Definitely, I would say that one of the challenges of migrating out of the mainframe is that business is not really seeing an immediate benefit to the top line into this mainframe migration because usually, the mainframe migration is happening with a like-for-like mindset. Business is pretty neutral and a bit cold when it comes to these modernisations that are more pushed by the technology department because of the cost pressure. I would say that unless there are specific cases where business may be interested in modernising a proposition in a certain area, big investment, big payment, big cards. In this case, this business is supportive of the fact that you are missing critical functionality. You target the migration more like a migration and enhancement and not just a like-for-like. In that case, you can get prioritisation. If you don't have this kind of premium for the business, then it's very hard to prioritise mainframe modernisation over the rest of the portfolio, which is more focused on the digital and assisted channel enhancement, on data and analytics and on market technology.

[00:27:22] What critical functionality or business case would get attention and prioritise mainframe migration? What would typically make a CFO, who doesn't want to do this due to a lack of business case, see a link to modernisation?

AC [00:28:07]: Let me say that the business case makes sense also for the like-for-like because typically, when you get out of the mainframe, you can get rid in one shot of licence cost like A&N, but also of the mainframe cost like MIPS. Usually, you may have a business case, then you can argue on the break-even, but you can have a break-even in 5-7 years. The real element of differentiation is that you mount on top of the savings, also an element of top line, so an element of revenue because you take the chance of the modernisation to say, I missed this functionality, I don't know, specific management of investment product or you miss specific capability in servicing the customer. In that case, you can blend the cost saving deriving from the change of technology to the revenue uplift that you have because during the migration, you take also the chance to enhance. In that case, it doesn't… It's not anymore a like-for-like migration, but it's rather a migration and enhancement use case. It's a bit different use case.

[00:29:45] It seems that A&N has a decent competitive moat, given its speed and related inverted index. How has that trended today? Are there comparable mainframe or alternative solutions in the market today where that speed isn't quite the differentiator it once was?

AC [00:30:12]: You mean speed in terms of throughput, or speed in terms of transaction processing or speed in terms of time to market and let's say a pretty new project?

PA [00:30:25]: Throughput because I believe that's where A&N really excelled, and that's what helped it revolutionise the industry. Is that still a competitive moat today?

AC [00:30:47]: If your point is more about, let me say, the capabilities of getting out of the mainframe, there is a lot of appetite in the tech department to get out of it. I'm not sure whether I'm capturing properly your question. Can you please maybe elaborate it better?

PA [00:31:17]: A&N is a language that works on top of the mainframe solutions. Given that Software AG is A&N's supplier, do you see a problem with the company's software, or is this more a problem that the hardware that it runs on this mainframe is just being phased out?

AC [00:31:49]: In terms of advantage, I wouldn't say that the advantage is coming only from A&N, but the advantage is coming from A&N over the mainframe because actually the reliability, the resilience, throughput, the scale is actually granted by the underlying hardware. What is the actual advantage of A&N is that they were providing 20-30 years ago, quote, unquote, modern development pattern and a modern database technology when back then there was only the rigid Db2. A&N is more like a facility in the so-called computed assist design rather than an element of differentiation when it comes to performance and throughput.

PA [00:32:41]: Looking at that 20-to-30-year-old database architecture, is demand for that style of architecture still aligned with A&N today, or are other technologies being prioritised?

AC [00:33:00]: You mean whether the other technology on the market have a better or worse approach vs A&N?

PA [00:33:07]: Yes.

AC [00:33:09]: Definitely, it's much better because the technology evolved in the last 20-30 years. Basically, I would say that the pattern of development in microservices is the demonstration that you cannot have just a one-size-fits-all database technology, but you can have database technologies that are more fitting for the use case. Also, A&N is not event-driven architecture, which is enabling natively near real time. I would say that definitely, there are in the cloud-native architecture, much more benefit in terms of application development, in terms of scalability. Let me say that A&N cannot be considered comparable to any modern technology that have been developed in the last 3-5 years.

[00:34:11] You mentioned that AI assists with the migration by documenting the code and helping to translate it over to more modern languages. Could you quantify this? Previously, how long would a migration take without AI vs how long does it take with AI now?

AC [00:34:46]: AI is proving to accelerate the time to market up to 50%. Basically, you can get double speed compared to the typical way of working. I expect that if you include also the unpredictable time for the reverse engineering, this can reach up to 60-70% in my experience.

[00:35:12] If we start seeing more referenceable cases more frequently where AI cuts migration timelines in half, even without a strong top-line business case, is that shortened timeline enough to convince banks to go towards migration?

AC [00:35:41]: Oh, definitely, yes, especially because, as I mentioned, the overall migration process becomes more affordable, but also, I would say, more sustainable in terms of time that included this. Definitely, yes.

PA [00:36:01]: You mentioned affordability. What cost reduction do you think AI will bring? Regarding timeline, you said it was about 50% shorter, but you pay for the AI functionality as well.

AC [00:36:22]: The cost of AI is a change cost because you basically consume tokens to perform the actual code migration, code conversion, but it's also true that you pay tokens instead of humans. There is a fair 20-30% efficiency between the human labour and the AI agentic development. I would say that at steady state, you may expect a reduction of 20-30% of cost because of the reduced infrastructure cost. While for the cost of change, I expect a 30% reduction between using only human developers and employing generative and assisted software development.

[00:37:30] A&N has committed to ongoing support and innovation until 2050. How do you expect customers to plan their long-term mainframe maintenance? Is there demand to last until 2050?

AC [00:37:46]: I think that A&N made also a marketing move to somehow create a bit of consensus on the fact that the technology, also looking at the significant revenue that they had in the last few years is something that is running the core of the business. Hence, it's so mission-critical that deserves a strategy that is targeting 2050. If you want to have my opinion, on one side, there is a marketing move. On the other side, with announcement of the 2050 strategy happened before the raise of agentic software development cycle. I think that now they're also pretty aware that with AI, their 25-year time horizon will definitely be shorter.

[00:39:08] You mentioned that you expect spend with Software AG to be stable to decreasing. Could you quantify that? Regarding existing customers' A&N spend, should I be thinking of a 0-5% annual decrease, or different numbers over the next 2-3 years, probably up to 2030?

AC [00:39:46]: Up to 2030, yes, it's going to be three years. In three years, you are… If you start now, you would be in 2-3 years at around about 20-30% of the journey completion, which then reflects on the contract in a 20% reduction at steady state in terms of income that Software AG will see for A&N.

[00:40:26] What percentage of the customer base do you think could be launching migration projects away from A&N or away from mainframe in the next three years?

AC [00:40:38]: You mean the same percentage but applied to the overall mainframe?

PA [00:40:43]: No, the number of customers that are likely to start projects to move away from Software AG and away from the mainframe.

AC [00:40:55]: The question is, what's the relative percentage of customers on mainframe and A&N that would start this migration?

PA [00:41:03]: Yes. Could we be looking at 50% of banks using A&N starting a migration process in the next three years?

AC [00:41:12]: I would say, yes, that 50% is a fair assumption, maybe also conservative from certain aspects considering the growth of LLM capability. I would say it's fair, a bit conservative.

PA [00:41:32]: What could that be looking at? Could that be 60%, 75% or more?

AC [00:41:39]: I think that 50% is a reasonably fair expectation.

PA [00:41:54]: In the next three years, we're probably going to see about 50% of the customer base looking to start migrating away from the mainframe and, therefore, away from Software AG. That comes with a decrease in spend of about 20% over 2-3 years after that migration starts. Is that a fair summary?

AC [00:42:17]: Correct. It's a fair summary.

[00:42:21] You mentioned looking at Aris more favourably as it doesn't get bogged down with the mainframe. What's your assessment for its spend going forwards?

AC [00:42:43]: We are mainly targeting to keep it stable because my consideration regarding the risk, opportunity is depending on something that Software AG didn't do yet, which is providing a new scheme or a new capability to their offering for that product. I believe that we have a fair chance that in the next couple of years, Aris will stay stable because you are not going to increase it. In case in the next 2-3 years, they will be really increasing the value of the proposition and of the product, introducing agentic they have a fair chance to grow. Considering the current situation, I see more like a stable usage. If they don't do anything, it is stable in the next 2-3 years, I expect that in 2-3 years, people will start migrating out of Aris too. That's why it's important that Software AG understands that for Aris, they have a chance to convert the AI wave into an advantage for them. Thus, they need to timely move in order to get out on the market with a credible proposition because I see already a lot of new player entering the market with a fully AI-native proposition.

PA [00:44:44]: Is that the number one functionality for Aris to be an AI-native solution, or are there any other capabilities that you are looking for?

AC [00:44:53]: Aris is a system of record. It is a system that allows you to have a static view of the process and have some mining functionalities. This was built 10 years ago when AI was not on the market. I believe that the capability of advisory on process improvement, leveraging on the reasoning capability of AI are definitely higher than what a human could be. That's why only having the product not based on AI will be shortly be considered as a weakness for the client. That's the reason why I believe that over the next 1-2 years, I see stability where a lot of customers would expect Aris to come on the market with a new proposition. If they won't do like this, they will see also new challenger coming in and gaining market share at that remains of Software AG.

PA [00:46:20]: Would you expect most banks to take a similar approach, maintaining the spend with Aris and not really wanting to expand in case there's a new proposition and it doesn't release a new update for the next few years?

AC [00:46:37]: Yes.

[00:46:41] I've read about business process mining and process intelligence coming closer together, with customers expecting them to be unified. What is your take on that?

AC [00:46:59]: You mean my take on the business of AI with process mining?

PA [00:47:06]: Yes, with process mining, and for instance how Celonis called it process intelligence, adding AI to this layer. What's your take on that story?

AC [00:47:18]: I think that Software AG has a competitive advantage of being on the market since a long time and many large organisations have that process must on Aris also for compliance reasons. Basically, they have the advantage of being the system of record for processes. Now once you have this amount of data that is crucial in the era of generative AI because generative AI gets the best when they have data and context. Now there are two possibilities. Either you just consume the information that you have within Aris, outside Aris, you consider Aris just as a system of record, you upload the data and you use them somewhere else. This would be more like favouring stability basically they keep their market share and customers will be using it. The second chance is that customer will start seeing that there are tools that are able to migrate automatically the information from Aris and then use this information in a much more fruitful way in terms of productivity exits. In that case, then Aris is going to have a big risk that basically the customers will be in a much easier position to migrate out of Aris.

[00:48:52] I read that there were fears that Aris could easily be disintermediated by generative AI, not by a specific software or solution, but just by running generative AI through systems. Do you think that's a credible risk?

AC [00:49:17]: The risk of losing market also in Aris?

PA [00:49:21]: Yes, the market share loss from Aris, but not to a bespoke software, just to generative AI usage.

AC [00:49:31]: You mean without any alternative products just like you manage in a custom way, process data and you use them through an LLM without any third-party software?

PA [00:49:44]: Exactly.

AC [00:49:46]: I think that considering the fact that those departments that are running Aris are not expected, I don't see really likely that they can get rid of third-party products and are dealing with the process data in a sign-ops (? 50.07). I see more likely that they will find another player rather than dismissing Aris or paper and pencil process where they get in the process information and then they use LLM because you have also a topic of process life cycle where you need a tool that is actually allowing you to make all the edits, make them auditable. There is a business around the life cycle of the process, and this cannot be performed through an LLM. The LLM is more like which insights, which advice you will get on making the process better, making that process more progressive.

[00:50:50] What other process intelligence or process mining solutions do you think customers might be looking for? Is there anything that ticks that AI criteria for you today, or is this still a wait-and-see for customers?

AC [00:51:11]: No, it's currently something very active on the market. On the market, we have many players that already went out with their proposition and (inaudible 51.23), think about UiPath, think about ServiceNow. From this standpoint, Software AG is lagging behind. That's why, if they do not react, they will be paving the same migration-out effect also in the area of Aris beyond Software AG.

PA [00:51:54]: Customers are staying with Software AG for the moment, despite competitors being there. Why do customers stay with Aris now rather than go into UiPath or ServiceNow?

AC [00:52:07]: That's my point because basically, it's either because these capabilities are now getting on the market. It's not that we have them in a few years. You have a physiological inertia where you keep your market share advantage because that Aris is having a market share advantage in current presence. If you don't do anything, people will start looking around for an alternative.

PA [00:52:40]: When do you think customers might start looking? Is that by the end of 2026, 2027 or 2028?

AC [00:52:48]: You mean in terms of still on process mining or more broadly?

PA [00:52:59]: Process mining, particularly alternatives to Aris.

AC [00:53:04]: How the market will behave in terms of how many new players we will see around?

PA [00:53:10]: When do you think customers will look towards someone else? You said that UiPath and ServiceNow already have those capabilities that customers want, yet the customers are sticking with Aris for the moment. How long will the customers stay until they meaningfully start changing away from Aris without any update?

AC [00:53:34]: 12 months.

[00:53:42] Looking at a general contract with Software AG, can pricing be increased to offset any level of customer churn from Aris, or the strip-back in mainframe and A&N spend you mentioned?

AC [00:54:06]: You mean what is the maximum increase that a customer would be accepting?

PA [00:54:14]: Yes. What would customers accept, or do you think it's realistic that pricing could be used to help offset?

AC [00:54:25]: I don't think that Aris is going to come back with a lower pricing. I expect more that they will keep the price constant because, of course, the price is a reaction of the market. I don't see in the near future market growth for Aris. If they don't get out with a compelling announcement of an AI roadmap, I see very likely that the projection of the revenue for Aris will be decreasing in the next few years.

[00:54:57] Could A&N's maintenance pricing be increased to offset any customer churn?

AC [00:55:07]: This is likely. This is exactly what IBM is doing on the mainframe. Basically, you start entering in a cash-cow mindset. You say, I'm going to have only this year to survive, let me monetise as much as possible what I have. I expect indeed that the primary result of having less client would be a price increase to the clients that have problems in getting out of it in order to compensate the churn at a broader level.

PA [00:55:42]: What price increase could be pushed through? Could we see a 10% increase in pricing?

AC [00:55:50]: I think that 10-15% is reasonable. Of course, they will sell it as more they get in future, they have a roadmap, they have a cloud-enabled capability that maybe you don't need, but they will try to put it or to sell it as a reason why the price is increasing without disclosing that actually the price is increasing because they have a reduction in the customer base and they're compensating the unit prices to keep the revenue up.

PA [00:56:20]: Would that increase happen just at renewal, or could that increase happen every year?

AC [00:56:27]: You mean in terms of why they would keep using?

PA [00:56:39]: More specifically, if contracts are renegotiated in 2027, is that when the price will increase 10-15%, or could it also increase in 2028 and 2029?

AC [00:56:53]: I expect between ’27 and I think it's going to be a phenomenon which is actually worsening over the years because, of course, the more they will see clients churning, the more they will be revising their pricing model.

PA [00:57:10]: Would that be with the same customer paying more and more every year, or is it only upon renewal that year that you pay 10-15% more?

AC [00:57:21]: I think that it's more affecting the contract renewal because usually, if you have a committed contract, you cannot have price increases during the duration of the contract unless they are driven by inflation indexation.

PA [00:57:32]: Is there anything else that you think is worth mentioning regarding Software AG?

AC [00:57:48]: No, I think we went through all of the topics and it's actually pretty interesting to see this effect. Definitely, I believe that AI is materialising more of a risk for Software AG they don't date, I would say, change of set-up and (audio distorts 58.06).

PA [00:58:07]: Thank you so much for giving me some of your time today. It's been an absolute pleasure, and I hope you enjoy your evening.

AC [00:58:14]: Thank you so much, and I wish you the same.

Transcription ends at 00:58:28 of the recorded material