AI in SAP: Why “start smart, not big” is the winning strategy

I’ve spent the past year immersed in the rapidly evolving world of artificial intelligence (AI) and its integration into the SAP ecosystem.

The opportunity is immense, but so is the confusion. Every week, I talk to clients who are both excited and overwhelmed by the pace of change, the proliferation of AI tools, and the relentless marketing hype. My advice? Embrace AI, but do it with your eyes open, and always start with a pilot project to prove the business case and test your approach.

AI in SAP: Hype vs. reality

SAP has been embedding AI and automation into its solutions for years, from process automation in S/4HANA to advanced analytics in SuccessFactors and CRM, there is a huge business value in these when adopted by users.

The latest wave, generative AI and “AI agents” like SAP’s Joule, promises to take things to another level. We’re now seeing tools that can accurately answer plain-English questions, automate business processes, and even initiate actions on your behalf, not just surface information.

But here’s the reality: in my opinion, most New Zealand businesses aren’t ready to go all-in on AI just yet. The technology is still maturing, as are the business models and pricing. For example, SAP’s Joule for Consultants offers a block of “AI credits” per user, it is therefore important to have clear understanding on how quickly those credits are consumed or what real-world usage will look like for your organisation. This is not unique to SAP, it’s a challenge across the AI landscape.

The data foundation: Clean Core, clean results

Before you even think about deploying AI, you need to get your data house in order. That means ensuring your SAP systems are up to date and your data is clean, accurate, and accessible. The old IT adage “garbage in, garbage out” has never been more relevant. If you don’t trust your data, you won’t trust the answers AI gives you. You’ll waste time double-checking everything, defeating the purpose of automation.

SAP’s “Clean Core” strategy is critical here. Make sure your systems are running the latest versions, ideally in the cloud, and that your data governance is rock solid. This isn’t just a technical exercise, it’s about building confidence in the results AI will deliver, both for your team and your customers.

Start small: The power of the pilot

One of the biggest mistakes I see is trying to “boil the ocean”, deploying AI everywhere at once, without a clear understanding of the business value or the risks. The smarter approach is to start with a focused pilot project. Pick a process or business function where AI can make a tangible difference, maybe automating invoice queries in finance, or using AI to summarise job applicants in HR. SuccessFactors, for example, is already seeing strong use cases in recruitment and onboarding automation.

A good pilot will help you:

– Prove the business case with real numbers (time saved, errors reduced, customer satisfaction improved)

– Build internal capability and confidence in AI

– Identify gaps in your data or processes that need fixing

– Understand the true cost and consumption patterns of AI tools

Once you’ve demonstrated value in one area, you can expand, confident that you’re investing in what works, not just what’s shiny and new.

Don’t wait for perfection – but do your homework

The pace of AI development in SAP is staggering. New features are being added every quarter, especially in cloud versions, and the roadmap is evolving fast. If you wait for everything to be “finished”, you’ll be left behind by competitors who are already experimenting and learning.

That said, don’t jump in blindly. Do your research:

– Study SAP’s published roadmaps to understand what’s available now and what’s coming soon.

– Take advantage of training and certification resources from SAP and its partners.

– Talk to others in your industry about what’s working and what isn’t.

Security and privacy: Know your boundaries

Finally, as AI tools increasingly connect with external data sources and third-party models, be mindful of data privacy and security. While SAP and its partners have robust controls, every integration point is a potential risk. For most businesses, the first step should be to master AI within your internal SAP landscape before reaching out to external data clouds or models and don’t ignore the benefits quickly achievable through the embedded AI and business process automation included in the latest SAP solutions.

AI represents a generational opportunity for SAP customers, but only if you approach it with clear eyes and a disciplined, evidence-based strategy. Get your data right, start with a pilot, prove the value, and then scale up. The future belongs to those who experiment, learn, and adapt—one project at a time.