There are '50 shades of clean core' for SAP customers

A large number of SAP customers are deciding how they can move from legacy systems like SAP ERP Central Component to the cloud and take advantage of innovative technologies like AI.

One of the key components of a cloud-based system like SAP S/4HANA Cloud is the clean core infrastructure, which relies on standard cloud processes rather than the heavily customized environments that are characteristic of ERP Central Component (ECC) systems. In the SAP clean core, a company's differentiating processes are minimized and can be handled in the SAP Business Technology Platform (BTP) development environment.

However, moving to the clean core is a complex endeavor, and SAP customers must rethink how they will handle or eliminate customized processes, which have often been in place for decades. Veteran SAP practitioners Michael Lemashov and Denis Malov of JDC Group, who work with SAP customers on a number of initiatives, advise taking a methodical approach to clean core development and a value-centered strategy for AI innovation.

JDC Group is an Atlanta-based SAP consultancy that focuses on services like S/4HANA migrations and implementations, developing SAP BTP applications, and SAP data transformation initiatives.

Lemashov, global head of cloud transformation and innovation services at JDC, focuses on machine learning, data science and AI strategy for SAP customers. Malov, chief architect at JDC, leads architecture teams in designing and developing AI, machine learning and other innovation projects for SAP systems.

What should SAP customers think about as they consider developing a clean core as part of an S/4HANA Cloud move?

Michael Lemashov: For the clean core, we say that there's not just one way to look at it, there's really "50 shades of clean core" for every company. It's not a final destination; it's a journey, and each company has to decide how clean they want to be. SAP suggests that companies need to be as clean as possible so that future upgrades are much easier and faster to do. This is the ideal, but the reality is that many companies have invested a lot in developing some functionality that was white space for SAP, so they have to find ways to make some alterations in a way it was done in the past and bring it to as clean an environment as possible. But it's not easy, especially when you've invested thousands of [development hours] to get that done. The reality is that companies are very complex, and there's a lot of demand for them to innovate and innovate every day.

Do customers need to move to the clean core before they can implement innovative technologies like AI?

Michael Lemashov

Lemashov: There's no way you can tell your CIO or CFO that you can't innovate until you clean up your core and do the upgrade, so you have to find ways that allow the companies to make innovations while they're also trying to come to a cleaner core and upgrade to S/4HANA. We like to say that you're trying to land the plane while you're still pouring the concrete on the runway.

A large number of SAP customers are still on legacy platforms like ECC, which are still usually on-premises. How should they think about getting in on technologies like AI? Are they looking at options like SAP BTP as an innovation center?

Lemashov: In the last few months we've done more AI-focused workshops, and a majority of the companies that did these workshops are still on ECC. They are due to upgrade to S/4HANA, but it's probably two years away, and they're getting demands from their C-level executives to innovate. So how do you do that while you're still on ECC and are looking to upgrade in a couple of years? BTP is a great option, and a lot of companies have started to look at BTP and are looking at replatforming their [SAP] PI/PO to [SAP] Integration Suite, where they typically stop.

But the reality is that you can do a lot of innovation on BTP, like building applications or building some AI capabilities. Then, when you do upgrade to S/4HANA, it's really easy; all you have to do is change the API calls on the back end. Instead of calling ECC, you are now calling S/4HANA. So this is not a waste of time, this is not something that will go down the drain, and you have to rework.

How should customers consider developing an AI strategy?

Lemashov: The picture with the customers is not black and white, it's not simple. A lot of customers are still on ECC, and whether they're looking to upgrade in the next year, two or three, a lot of them already have an AI strategy, and a lot of times this AI strategy is not necessarily SAP-centric. We've seen that a lot of customers over the years have been taking data from SAP, loading it somewhere in Snowflake or other things, and then using AI technologies from SAP competitors. A lot of these companies have fairly mature data science, machine learning and AI practices already. Innovation doesn't wait for anybody; if there's a need and a demand, companies will find a way.

But we've also seen in our workshops over the last few months, that a lot of customers don't even have AI strategy. And if they don't have AI strategy, most of the time there is no mature AI organization or practice. They're still looking around to figure out where they should make their bet: Is it with SAP AI technology, or with the hyperscalers? But for the ones that already moved on [from SAP], it's going to be challenging to bring them around and make them start using SAP technology.

What should customers be asking as they consider an AI and innovation strategy?

Denis Malov

Denis Malov: Definitely develop an AI strategy that focuses on value, but don't [wait in developing the] AI strategy. Obviously, it's a quickly evolving [technology], but in order to ensure your business continuity, don't disrupt too much upfront. Focus on the value. Don't think about [having] AI everywhere, but focus on where AI can create the biggest value.

Lemashov: The AI field is evolving rapidly, and we have these new, great shiny things coming out on a weekly basis. But we've observed that when customers lose their focus because of that vendor hype, their AI projects tend to fail very quickly. Some of the older LLMs may be perfectly feasible to do certain tasks or jobs, and they don't cost pretty much anything. So why go for something that's super new, super shiny and super expensive, when you can get things done with something that's already stable and been around for some time? If it brings you value and it's not costing you an arm and a leg, why not use it? Don't go for that hype, go for practical and look for outcomes that deliver value.

Jim O'Donnell is a news director for Informa TechTarget who covers ERP and other enterprise applications.