Presented by SAP
In an era where anyone can create an LLM, the real differentiator is not the AI technology itself, but the institutional knowledge on which it is grounded. Internal consultants and operational transformation partners cannot risk hallucinatory guidance when their recommendations impact integrated processes across supply chain, manufacturing, finance and other key functions.
“Established AI is non-negotiable because accuracy is not optional, such as when we execute million-dollar transformation projects in the SAP ecosystem,” says Natalie Han, vice president and chief product officer, AI gene at SAP Business AI. “Search-assisted generation technology and the ability to anchor answers in trusted enterprise knowledge help provide accurate code interpretation, best practice guidance and simple decision support. This is how we ensure true trust in AI-powered consulting.”
He adds that a fully grounded AI assistant such as SAP Joule for Consultants is of great value for production employ cases. SAP Joule has terabytes of institutional data that is constantly reviewed and updated, so your consultant can be confident that when using Joule, they are receiving the latest SAP best practices and methodologies while accelerating project execution.
“We save 14% in rework time, and consultants save 1.5 hours per day per user, which is huge considering how expensive consultants are today,” Han says. “Early adopters like Wipro estimated they saved 7 million hours for their consultants through manual operation.”
SAP Joule foundation
SAP Joule is as certified as any consultant, says Sachin Kaura, chief architect of SAP Business AI. The tool was born in 2023 when GPT graduates passed a simulated bar exam and created a buzz around LLM graduates’ ability to handle a lot of context. It is widely recognized that the SAP ecosystem, along with its associated ontology and taxonomy of domains, is extremely huge and can be very elaborate to navigate. The question arose: How can an AI co-pilot be used to navigate such a elaborate situation when it is actually embedded in the SAP ecosystem itself?
Sachin Kaura began experimenting with pioneering LLM models, subjecting them to the same certification exams that SAP consultants take. Initial results were penniless, but after extensive contextualization and a focus on delivering value to the partner ecosystem, Joule has consistently achieved results of 95% or better.
“Not only did we test from a data perspective, but we were able to work with all of our consultants to achieve what we call a golden dataset,” Han added. “It is non-deterministic, language-based, and completely consultant-driven. We partnered with the entire consulting organization to manually mark the gold data set across all products. This has become the foundation of everything we do even now.”
State-of-the-art indexing pipeline
Joule for Consultants is updated in real time. A state-of-the-art indexing pipeline introduces modern SAP documentation and makes content available to the model as soon as it is published, giving consultants confidence that each response reflects the most up-to-date guidance.
“This is pure engineering work performed by our analysts and data engineers, using many of the core SAP technologies,” explains Kaura. “We leverage SAP’s core business layer, document grounding services, and many purpose-built systems to stay up to date with current events in the system.”
SAP Business AI also operates at the management level, so it is not the task of just one team, but the priority of the entire company. They have built powerful internal partnerships with content owners across the SAP system – including SAP Learning, SAP Community, SAP Lend a hand, product teams, and consultant teams. Together, they continually update proprietary content such as SAP notes, knowledge base articles (KBAs), and other domain-specific guidance that reflect evolving SAP best practices.
All this means that Joule for Consultants can pull constantly updated data and provide answers in near real-time. This is the type of research that would otherwise take up a consultant’s hours. However, information obtained directly from the source provides consultants with the most up-to-date and reliable guidance, helping to eliminate early-stage errors that can derail a project months later when the scope has not been up to date with the latest capabilities.
Ensuring enterprise-class security
Han says SAP is building a product that is relevant, reliable and responsible. As a European-founded company, we take data privacy seriously, adhering to GDPR and other EU corporate regulations. At the heart of SAP Business AI is AI Foundation, an AI operating system that manages artificial intelligence with built-in security, ethics, and orchestration, using automation and intelligence to manage lifecycles, optimize resources, and enhance resiliency.
All LLM solutions used by SAP and its customers are powered by artificial intelligence that protects private and proprietary data from leakage. In addition to data protection, SAP also considers bias, ethics, and security at the enterprise level, with humans checking and balancing solutions on an ongoing basis.
“We have an enterprise-grade safety framework, as well as rapid injection and guardrail testing,” says Kaura. “The orchestration layer built within the AI Foundation anonymizes input and moderates it to prevent malicious content. This ensures that the output we give to our customers is relevant to the SAP ecosystem and the domain they are asking about, not just a general LLM glut. This set of tools, from the framework layer, to the application layer, to product standards, as well as very thorough testing, is critical to securing our product. Then, and only then, can it reach our customers and partners.”
Pushing the boundaries of Joule for consultants
“We are just scratching the surface of what LLM and agentic AI can offer,” says Han. “Access to knowledge is just the beginning. We will have a much deeper understanding of customers’ SAP systems and will be able to help them implement and transform their journey. The product team and our engineers are working to make the tool more transformative, able to extract more insights, connect to customers’ systems, and understand and optimize their processes, including code generation and support for customer code migration.”
The next step is to apply a second layer of primer. SAP’s customer base is huge and its partner ecosystem has implemented countless business scenarios. The first milestone was embedding Joule in SAP’s institutional knowledge; the next is to layer on each customer’s own context – historical system data, process designs, implementation plans and internal documentation. This allows Joule to move from being SAP-aware to being customer-aware, providing guidance that aligns with how the company actually operates.
“Think of it as basing your knowledge on your SAP knowledge, which will give you more accurate and relevant guidance,” says Kaura. “Information that might otherwise be lost may be in Joule for Consultants. Our system processes it and ensures it reaches you in the right way and at the right time.”
With this expanded approach, Joule can tailor its guidance to suit the consultant’s role – whether he or she is working as an architect, functional consultant, or technical consultant.
“We provide the information needed for a specific customer configuration,” explains Han. “Then we will be able to not only answer general questions, but also answer their specific configuration. From there, it’s just a step to generating more insights and taking more actions.”
Sponsored articles are content created by a company that pays to publish or has a business relationship with VentureBeat and is always clearly marked. For more information, please contact us sales@venturebeat.com.