SAP’s CPO Gives the Insider Track on AI with BTP

⇨ Generative AI adds a dimension beyond just large language models (LLMs) and machine learning.

⇨ SAP’s role is to put AI technologies into specific business domains where its customers are doing the most work.

⇨ The foundation to SAP’s evolving AI strategy is its SAP Business Technology Platform.

Across the SAP ecosystem, AI may be the most buzzed-about technology. Companies are hurrying to infuse AI into finance, data management and many other types of solutions. Yet there are still plenty of questions surrounding AI as it pertains to SAP users: How mature are these AI technologies? Where can businesses get the best ROI on AI investments? How much of the AI hype can businesses actually expect to become reality? And how can privacy and copyright regulations be ensured, especially around critical business data?

To help answer these questions, SAPinsider sat down with Michael Ameling, SAP’s chief product officer and executive vice president for the SAP Business Technology Platform (BTP). He highlighted key trends, focus areas and considerations related to automation and AI which Insiders should be aware of to fully “hatch” the tech’s potential.

 

More than one basket of eggs

This tipping point between hype and reality occurs when a new technology becomes commodified. If you enjoy using something in your private life, you also want to have it in the enterprise world. For instance, when OpenAI’s ChatGPT tool became part of the broader technological ecosystem for users in their personal life, they wanted to apply that same functionality to their professional roles as well. What also fuels the enthusiasm is the fact that Generative AI applications enable anyone, regardless of their technical expertise, to engage with and witness artificial intelligence in action.

Generative AI adds a dimension beyond just large language models (LLMs) and machine learning. Generating text, images and sounds quickly at enterprise-grade quality is likely to forever change the way people work. As of 2024, there appears to be enough adoption of GenAI in the consumer space that means AI is here to stay. So how do users keep up?

One option is that users spend time themselves getting an understanding of large language models, their underlying technology and their associated use cases. However, this requires a significant time investment.

SAP’s approach, Ameling explains, is to provide customers who have an understanding of the basics on GenAI with concrete use cases specific to business. By cutting through the noise and delivering specific GenAI use cases, SAP customers can avoid getting lost in all the new technology trends and find a hand-picked solution that can deliver real business value. As in the consumer space, it’s important to keep the entry barrier low and seamlessly embed AI capabilities.

“We don’t know what the dominant large language model of the future will be or if there will ever be just a single one. There are different technical approaches and there is no one-size-fits-all solution. This is why we have a multiple vendor approach. We partner with OpenAI, Aleph Alpha, Cohere, Anthropic and others so that we can make use of the large language model that best fits a given use case,” says Ameling.

This benefits users, not just from a feature perspective, but also helps keep costs under control. As with any emerging technology, AI can be expensive. Having a wide array of vendors to select from can help organizations align their business goals and budgets to find the best AI solutions to maximize ROI. At the same time, customers can use different models for different scenarios and manage them centrally.

 

How SAP supports AI technology

SAP’s role in the emerging AI space is to put AI technologies into specific business domains where its customers are doing the most work. For instance, new AI technologies in finance settings are now able to match invoices after training on hundreds of documents. Before, this training required thousands or even millions of documents to achieve the required accuracy.

SAP aims to help all their customers leverage AI capabilities independent of the SAP solution they are using. They all have different sets of SAP and non-SAP solutions, but they share the ambition of making the most out of their SAP landscape.

SAP’s stated goal is to offer an open technology platform that allows users to leverage AI for all kinds of different business use cases. Many of these use cases are built into SAP’s portfolio to provide customers with immediate value. However, according to Ameling, SAP wants to support customers with sustainable business success, so legacy systems won’t be able to access the full breadth of these capabilities, as older systems often do not fulfill the technical prerequisites for more advanced options. Over the last couple of years, Ameling adds, SAP prepared for these kinds of scenarios by promoting a clean core strategy.

The foundation to SAP’s evolving AI strategy is its SAP Business Technology Platform. Thousands of users are already leveraging its capabilities to deliver AI, complementing their SAP landscapes. SAP BTP can integrate with any system, providing access to business data and datasets, helping users make the most out of using LLMs. While fine tuning adjusts the behavior of LLMs, the VP says, SAP’s vector engine capabilities make it “even simpler” to use data for embeddings to enhance the model’s knowledge.

“We will focus on business AI scenarios like in procurement and finance. On SAP BTP itself, it’s in analytics. People use AI capabilities so that charts can be generated. We will have coding capabilities and a lot of additional very specific use cases where we want partners to build on top of what we offer. When it comes to SAP S/4HANA transformation, we know there are a lot of custom coders out there. We partner with them to make it easier for customers to transform their IT landscape. AI is very promising for transforming the current system at scale,” says Ameling.

Ameling adds that openness is key to SAP’s AI strategy. SAP is not a closed environment – in fact, as he reminds us, it is one of the largest open-source creators and consumers in the world. SAP provides open connectors to facilitate innovation, while also ensuring that AI solutions are fully reliable, fully secure and available in all relevant markets.

“For businesses to trust generative AI, they need to be sure their data is handled safely and confidentially,” says the VP. “Let me be very clear here: No customer data will be used to train external foundational AI models.”

This is why, he notes, that SAP BTP provides AI capabilities for development and administrative roles to equip them with the tools needed to efficiently build and run business solutions infused with AI.

 

Action, not reaction

The use cases for AI seem practically infinite, but businesses need to be realistic about where they can actually apply this technology today to generate real value. As of early 2024, the AI space is moving beyond just being applied to certain scenarios. The next level of AI to “break out” is going to provide computer systems that are acting, rather than reacting.

Imagine a system that proactively informs a business if it is about to exceed its P&L or if there are certain demands on its supply chain forthcoming, so the system recommends or even places an order to fill a gap. This is the next trend to watch for in the AI space.

 

The “aha” moment

To make the most of emerging AI technologies, AI leaders need to get familiar with the technology to become true eggheads. This goes beyond just reading about potential use cases – leaders need to dig into these technologies themselves, get engaged with the tools and discover first-hand what is possible.

“Just try it out. There needs to be an ‘aha’ moment. Generative AI will change the way developers work, but if you don’t get in touch with it and try it out, you won’t get that moment. As soon as you get this ‘aha’ moment, you won’t be able to stop. This is where it gets fun. This is when it changes how you work,” says Ameling.

It is vital that leaders, developers and business users reach this “aha moment” so they can understand the possibilities that Generative AI will give them. Once this moment occurs, users won’t spend time on tasks like finding how to integrate the next API. Instead, they will work to create even better prompts to achieve the full potential of Generative AI and make full use of its low entry barrier.

Another key piece of the puzzle is equipping businesses with the right partners and the right technology to be successful. Users should evaluate all potential partners and solutions that meet their specific needs before getting too far down the road with Generative AI.

One last piece of advice Ameling imparts is to not fear failure. As with any new technology, progress will come in fits and starts. Users shouldn’t be discouraged if they do not see the exact results they were hoping for immediately. Experimentation is a key process in achieving value from AI.