SAP Challenges Value of ‘Best-of-Breed’ Apps in Enterprise AI

When executive Muhammad Alam stepped onto the Fountainbleau theatre stage in Las Vegas earlier this month to give his opening keynote talk at SAP Connect, I half-expected to witness a spectacle. After all, I’m at a tech conference where AI dominates the agenda, so why wouldn’t there be a flashy performance, brimming with pagentry, to hype the announcements? But after two hours, SAP’s address felt different—almost pragmatic and deliberate—and that was the point.

Typically, companies will use these moments to sell their vision of the future. But in the wake of the much-debated MIT AI study, some tech firms are shifting away from further hyping up the technology and instead focusing on educating their customers. Indeed, some view AI as over-hyped, being pushed onto everyone while sidelining concerns and feedback from businesses on how to use it effectively to deliver real results.

Instead of showboating, SAP took what appeared to be a more sensible approach, presenting AI as a strategic business tool, breaking down its impact into five specific innovation patterns, and highlighting its practical enterprise value. It was heavier on demonstrating how AI could solve real business problems across different departments, versus spending time glamorizing its announcements, which were largely new Joule agents.

“There are a lot of [experts] out there making a lot of headline-grabbing predictions on how the world of business applications will change in the age of AI,” Alam said in his remarks against a backdrop of news articles and quotes from Salesforce’s Marc Benioff and Workday’s Carl Eschenbach. “Some say it will lead to the end of SaaS, while others equally vehemently argue the other way. If you take a step back, though, it’s not hard to spot the correlation between the prediction and how accretive they are to the valuation of the predictor.”

Alam offers this counterargument, saying that future business applications will depend on a seamless integration of apps, data, and AI to provide maximum value. The absence of this “flywheel” could result in significant costs and missed opportunities. He also contends that AI with the broadest context will provide the deepest value to companies. It shouldn’t be limited to one or two domains, but all of them, while also being “deeply business process and role aware.” It’s here that he argues best-in-class tools will be advantageous for organizations over best-of-breed: ”Not just a nice to have, but a must-have in the age of AI.”

The Best-of-Breed vs. Best-of-Suite Debate

“When we say that we think that best-of-breed will struggle in the age of AI, it’s based on a very simple construct,” he tells me in an interview. “First, there is a new layer that commands value that, by definition, as the values shift up, the layers below will become more commoditized. Our point of view is that the core application layer will become more commoditized.”

In other words, SAP believes that having the most feature-rich application is no longer a competitive advantage. Instead, with AI now becoming the highest-value layer, tech providers must demonstrate how AI will work intelligently across all apps.

The best-of-breed versus best-of-suite discussions are about capability versus integration.

He explains further that it doesn’t make sense for CIOs or other enterprise leaders to pick specialty applications, such as buying human capital management (HCM), procurement, or employee resource planning from different vendors. Doing so requires businesses to take on the responsibility of building the needed integrations. “Those don’t come out of the box—for just the transactional piece,” Alam argues. “And I’ve spent a significant amount of money doing that, just so then the transactions can flow. So you have customer information from A to B, supplier from here to there. That’s a significant amount of work for what value? Questionable, because the value layer is shifting up.”

Another concern he raises is the data model, which might not harmonize across the different apps a business might use. Once the integration is complete, there’s still a need to extract the data, place it in a data lake, and make it consistent, compatible, and useful across the entire organization. And then analytics must be added on top. This can be an expensive process. “How does AI come to fit on top of it?” Alam wonders. “That is really the value layer…but at the commodity level, you can get applications A, B, and C without compromising capability, but pre-integrated so that you can focus on high-value AI [and] you don’t have to spend time harmonizing the data layer, why wouldn’t you? This would just make common sense for you to go do so you can shift your efforts to the AI layer.”

What Alam is essentially saying is that enterprises incur significant opportunity costs when buying software from multiple vendors, particularly around integration and data management. He tells me that 15 CIOs and CEOs from SAP’s largest companies have asked why they shouldn’t build custom solutions rather than buy off-the-shelf offerings. “My question back to them was, just because you could go build a procurement solution on the fly, does it make sense for you to go do that? You’re in the oil and gas industry. You have a lot of things to do. You generate a procurement app—then what? You’re going to integrate it to finance, you’re going to integrate it to this, and you’re going to maintain it forever, and you’re going to roll it out, and you’re going to figure out what the data model is? The lifetime cost, not the value of being able to do something…is what I think the industry is glossing over.”

He concedes that there are times when building is more opportune than buying, specifically if you’re already maintaining the apps. “Go generate an app where you have a power app or a low-code, no-code—kill those, because why do you need that still?” he declares. “Go generate an app where you were doing custom development anyway, because…those you were going to maintain anyways, but now you can generate them far quicker, but the lifetime [total cost of ownership] still stays the same.”

Rebecca Wettemann, CEO and principal analyst at Valoir, suggests that the software industry oscillates between “best-of-breed” and integrated suite approaches. “We get to a certain level of advancement and innovation across the different pillars of a suite, and then something else new comes along,” she tells me. “That becomes the new best-of-breed innovation. We see it time and time again in the software industry. We saw it with niche cloud players—now everybody’s a cloud player. We see it in the CRM space—all of a sudden, marketing automation was best-of-breed that everybody needed. Now, it’s naturally part of a suite. You can see it across sort of every enterprise software stack.”

Are Walled Gardens the Answer?

That thinking may be contradictory for some, especially since workers may prefer the best-of-breed approach. Is SAP advocating a return to the Microsoft playbook of a closed, integrated ecosystem in which companies are devoted to a single system and waiting for IT approvals before updating? As Alam frames it, this approach may simply align better with SAP’s customer base, which includes large, multinational, risk-averse organizations, operating under strict regulatory and fiduciary scrutiny.

If I have a Mac, and [an iPhone], and I have an iPad, everything just works seamlessly together. I don’t have to worry about syncing anything—and it works. I don’t carry a camera anymore. I don’t carry a GPS anymore. It is just a better experience for me to have that so we see that in our daily lives, every day, and this is the analogy of it in the business world.

Nevertheless, he acknowledges that some companies may prefer not to be locked into a single vendor. “We need to earn every single workload,” Alam responds. “Still, what I’m not saying is you should just come to us because the integration is there. That’s the old best-of-breed versus best-of-suite. To me…it’s best-of-breed-as-a-suite. You shouldn’t need to compromise on any capability. You get integration out of the box [and] the harmonized data model so that you can focus on the value. So hold us still accountable for it.”

He argues that true walled gardens can’t—or wouldn’t—exist in AI, distinguishing today’s environment from the closed, defensive ecosystems Sony and Microsoft built decades ago. “Anybody could put up a [Model Context Protocol] server in front of an SAP application today,” Alam remarks. “There are two things that will differentiate AI over the fullness of time: One is accuracy (how precisely the AI can answer or perform tasks), and the other is performance (how fast the AI can generate results).”

The walled gardens of yesteryear are more challenging to implement thanks to MCP, Agent2Agent, and other interoperability standards. Because of this, Alam disagrees with the Microsoft comparison, saying that SAP’s appeal is its knowledge graph, fine-tuned models, and deep understanding of business context, which spans across ERP and finance to supply chain and manufacturing, customer experience, human capital management, procurement and business networks, sustainability, and other sectors.

A Pragmatic Keynote in a Hype-Filled World

Among the announcements made at SAP Connect were the debut of the SAP Engagement Cloud and 16 new specialized Joule AI agents. Compared to similar events put on by its peers, the news felt uninspiring. In a way, it felt better if this had been conveyed over email. As it turns out, grasping what SAP was laying out this week required a broader view than I first realized.

“Companies and CIOs today have a very hard job because they’re getting all of this glitz from everywhere,” Alam explains, noting that he found my viewpoint perceptive. “And not just that. There’s so much out there, and then the boards are telling them, ‘show me value,’ and it’s hard for them to figure out how to make the right decisions on what I go do?”

The goal of his keynote, he reveals, was to logically lay out a point of view on how companies should think and approach AI. It wasn’t about spending more on SAP—though he wouldn’t argue against it—but guiding business leaders inundated by all the AI news coming at them. “You could pick things that you’re going to end up paying for for a very long time, from a total cost of ownership perspective over the lifetime,” Alam says.

He claims he also wanted to provide a “frame” for companies on how to think about AI, especially as more companies announce AI-related layoffs—though he intimates that if you look closer, the real reason for the separation isn’t entirely attributed to the technology. Alam attests that SAP is best positioned to provide this business framework because it’s “a company that arguably understands businesses better than probably any business app company out there. At the breadth of what we do—you can argue that Oracle is probably the only other one that has the breadth that we do—is how to frame and think about AI, because it has to go with humans. You have to partner with it. And all the things that I talked about make the human smarter, more efficient with a bunch of tools.”

How Joule Reinforces SAP’s Vision

All of this was reflected in Alam’s keynote, which was surprisingly pragmatic for a tech event centered on AI. Rather than selling a utopian vision of a fully automated future, SAP leaned on relatable case studies to illustrate real-world applications. This approach, again, likely influenced by the MIT AI study, emphasized tangible business value over hype. Even the new Joule AI agents, though given their moment in the spotlight, often felt secondary—almost as props supporting the broader message rather than headline-making innovations in their own right.

With Joule, we’re just not providing you a blank canvas…we’re providing you [with] out-of-the-box agents and AI assistants, because we understand the depth of the processes in the work that you can get started tomorrow, as opposed to you having to think about which agent to build and how to go connect.

Announcing fifteen specialized agents as the centerpiece of an event of SAP Connect’s caliber may leave people wondering why even have a gathering at all. Yet, in SAP’s framing, these pre-built solutions were more than novelties. If they do what’s promised, they’ll deliver immediate value, integrate with core business processes, and reduce the need for custom development.

“The best way to land this in an enterprise is to connect it to the people that will use these tools to make them better,” Alam states. “And as the tools become much better, more complete, and more reliable, then you can connect the tools to get to autonomous execution. And that’s really where we think we can play a unique value because it’s—outside of one other company that I can think of—nobody else has this end-to-end business process context understanding of the business processes [and] the industries other than we do.”

He stresses that most companies probably don’t want to waste time and resources building complex AI tools from scratch, even though they technically could. Instead, he thinks they want ready-to-use AI solutions that work immediately within their specific business functions. In other words, just because you can build something doesn’t mean you should. Most companies want practical, immediate solutions that can solve real business problems without requiring extensive custom development.

That appears to be what SAP aimed to do at SAP Connect.

“It was very intentional, because I believe CIOs and the customers are in a very tough spot,” Alam admits. “Just look at the [OpenAI Dev Day] announcements…that were happening in parallel, and, oh, I could go create my own app, and then the CEO, who doesn’t know any better, is probably going to send a note to the CIO and say, ‘Why are you asking me to sign this business apps contract? Can’t you just go recreate?’ So I think a lot of it is education and clarity we need to help create for our customers.”

Still, despite his pronouncements, will SAP’s message move the needle and convince organizations that its “best-of-breed-as-a-suite” approach is the best way forward when it comes to AI? It might resonate with executives from slower-moving companies or those in regulated industries. As Alam notes, CIOs are flooded with AI news and hype, so the question remains: Has SAP demonstrated that it can genuinely drive impact and serve as a reliable guide for enterprise AI transformation?

As our 45-minute conversation wound down, I asked Alam whether he subscribed to the “digital workforce” vision promoted by Salesforce and Microsoft, his former employer of 17 years, and where he led Dynamics 365. He was candid in his response, questioning whether those companies truly “know the business.” A supply chain transformation conversation between a CEO and Microsoft’s Satya Nadella, whom Alam describes as one of the smartest people he’s ever worked with, he noted, would sound very different from one with SAP’s chief executive, Christian Klein.

“What [Nadella] is selling is, ‘move your supply chain systems to Azure, and move all your data to Fabric, and this is your supply chain of next generation.’ What [Klein] is talking about is, ‘how do you run manufacturing, demand planning, and asset performance’—that’s core supply chain, if you will. So, for somebody who doesn’t understand the business to say ‘we’re going to build the next Frontier Firm,’ is…it’s a lot of empty shells out there…but what we’re doing is we’re building it from within to say, ‘with this assistance now, this autonomous execution on this process…these are real things. We’re not talking about a frontier blah. We’re actually building it from the ground up…and doing the hard work.”

Disclosure: I attended SAP Connect as a guest of the company, with my flights and hotel costs covered. SAP did not dictate the contents of this post. These words are my own.

Featured Image: An AI-generated rendering of SAP executive board member Muhammad Alam, who oversees product and engineering, responding to media and analyst questions at an event during SAP Connect on Oct. 7, 2025. Credit: Ken Yeung/Google Gemini