German software giant SAP has pitched a business suite set of integrated applications as the foundation that companies need to succeed in artificial intelligence (AI), claiming that fragmented data landscapes are the biggest barrier to advancement.
With a business suite of “best of suites” bundles that integrate data, processes and capabilities in areas such as sales and finance, SAP is directly challenging the “best breeding” model where organizations gather fragmented applications from different suppliers.
Stephan de Barse, SAP's global business suite leader, argued that the best breeding models are not suitable for AI. “The fragmented application landscape makes it difficult to create fragmented data that loses business context when drawn into a data lake for analysis,” he said. “If you don't modify the data, there's no good AI.”
According to De Barse, this fragmentation means that an organization spends around 80% of the time and effort it takes to manage its applications and data, and only 20% remains to deliver business value.
Instead, SAP promotes the “flywheel” concept. This concept brings AI, data and applications together in a noble cycle. SAP believes that by integrating data and applications within the business suite, AI can maintain the context needed to generate meaningful and actionable insights.
To explain this future, SAP demonstrated how its AI co-pilot Joule becomes a central point in user interaction. “Users will no longer be logged in to the application. Instead, Joule is the front-end and coordinates the workflow across multiple parts of the SAP suite,” says De Barse.
For example, tasks such as responding to new customer orders with manual interactions with the manufacturing and financial systems can now be handled via a single conversation request to Joule. This will take the necessary actions in the background software.
SAP is also developing AI agents with a variety of operational capabilities, from helping managers on-site to recover from power outages to helping less experienced staff make up for absent colleagues. De Barse noted that some customers have already begun to include these agents on their organizational charts.
“The first eight agents are already with customers. It will grow to 40 by the end of the year,” he said. “This development is growing very rapidly. This is a massive paradigm shift in enterprise software.”
Customers put AI strategy into practice
At the SAP Now Conference in Melbourne, several SAP customers shared how to use AI, balancing target projects with initial exploration with concrete results.
SA Power Networks (SAPN), a South Australian power distributor, uses AI to improve asset management and customer service. Travis Smith, SAP Domain Architect at SAPN, said the company's philosophy is to “go slowly to get faster” by ensuring that the right foundation building blocks are in place before scaling the project.
SAPN has developed a predictive analytics model for Stobie Pole corrosion using SAP's Business Technology Platform (BTP). With over 600,000 distinctive power poles across the state, this model helps SAPN move onto risk-based testing schedules.
“Which is the question? [poles] Do I need to schedule an examination? ” Muen Chen, data and decision science manager at SAPN, said corrosion is likely.
Smith and Chen provided practical advice to others on their AI journeys, highlighting the importance of centralized identity systems when using Joule in multiple applications. They pointed out that Joule struggles to interpret the data presented in the table and can require that the information be formatted in a different way in the source document.
Recently merged with Chemist Warehouse, Sigma Healthcare uses SAP SuccessFactors and BTP to automate complex HR processes. Chief Human Resources Officer Daniel Di Pira said automation previously saved 70% of the time it takes to generate around 1,000 employee contracts a month. This is a complex, state-specific process.
Meanwhile, other organizations are proceeding with caution. Chris Unes, Chief of People and Culture in Sydney, said the city is still in the “hype phase” of AI. “We can show you the business case and then we can investigate,” he said. He added that the technology is “not mature enough” to meet all expectations.
However, Youness sees clear possibilities for AI to draft complex documents such as skill frameworks. According to him, the key is to keep the human in a loop. “We still need human eyes,” he observed. “It's built for us, for us, through AI.”
Sean Black, who plays the CIO in Energy Queensland, said his organization uses machine learning and other AI technologies, and it's a question of when, not when, generating AI will be adopted.
However, the energy industry is highly regulated, so his organization focuses on responsible AI frameworks and target pilots. He argued that organizations should not treat AI as an afterthought, but should “design for AI” from the start.
“We're interested, but we were cautious about taking AI,” Black said. He predicted that just as AI became an invisible part of consumer applications, it would ultimately become an integrated, assumed component of business software.
