Anirban Ghoshal
Senior Writer

SAP to infuse IBM’s Watson AI engine into its entire portfolio

News
May 02, 20235 mins
Artificial IntelligenceGenerative AIIBM

The move is expected to help SAP exploit the natural language processing capabilities of Watson AI along with predictive insights, with the aim of jointly developing large language models and generative AI capabilities.

ibm watson on jeopardy
Credit: IBM

ERP software provider SAP on Tuesday said it is partnering with IBM to infuse the latter’s Watson artificial intelligence (AI) engine across its entire solutions portfolio, including SAP S/4 HANA, S/4 HANA Cloud, SAP Business One, and SAP Business ByDesign.

The move, which is expected to help SAP exploit the natural language processing (NLP) abilities of Watson AI along with predictive insights, is aimed at boosting the productivity of companies and accelerating innovation, especially across the retail, manufacturing, and utilities sectors, the companies said in a joint statement.

IBM Watson’s capabilities, according to the companies, will be integrated into the digital assistant inside SAP Start, which is a service that runs on every computer where an instance of an SAP system is started.

SAP Start monitors the runtime state of all SAP systems, processes, and instances along with reading logs, traces, and configuration files among other threads.

“With SAP Start, users can search for, launch and interactively engage with apps provided in cloud solutions from SAP and SAP S/4HANA Cloud,” the companies said.

The assistant inside SAP Start can also be used to run machine learning and AI to extract information from a variety of data sources and answer business queries, the companies added.

Typically, SAP uses the term digital assistant to collectively address or define all the digital assistants and chatbots in its portfolio, according to Juergen Butsmann, head of intelligent technologies in solution management for S/4 HANA.

“The Digital Assistant is the end-users ‘support center’ for ad hoc tasks, help, FAQ, or any other off-context inquiry. Whenever a user requests tasks or services (whether that be via commands, chat questions, or, in some cases, voice commands) digital assistants/chatbots can convey and interpret the user’s request, execute it efficiently in the system, and promptly provide a response,” Butsmann wrote in a blog post.

IBM Watson integration to aid back end automation

The integration of IBM’s Watson AI into SAP’s portfolio will not only help automate back end tasks but also aid in generating business insights, said Bradley Shimmin, chief analyst of AI platforms at Omdia Research.

“I think the integration will help SAP customers in creating a natural language assistant that is capable of taking programmatic action on backend services. This can help SAP customers maximize the value that SAP offers as part of its rich yet diverse product portfolio,” Shimmin said.

The other value proposition of the Watson integration, according to Shimmin, is the ability to surface enterprise knowledge from across large and disparate datasets.

“Experientially, these two capabilities when intertwined will allow SAP users to surface and take action on valuable business insights,” Shimmin said.  

Is SAP’s digital assistant mature enough in NLP?

Although SAP invests in its own research on AI, the reason to use IBM’s Watson to power its services is SAP’s lack of maturity in natural language understanding capabilities, according to Shimmin.

“Like IBM, SAP certainly has invested heavily in its own AI research. However, that research hasn’t focused on NLP and even natural language understanding,” the analyst said. “It is within these areas that IBM has created value, a value which SAP obviously feels will help them reach their goals with SAP Start more rapidly than if they were to develop this self-same technology in-house.”

SAP’s approach was an effective strategy given the demand for AI features in the market and the rapid evolution of the technology, Shimmin said. “Not every big software player needs to build its own ChatGPT or Bard in order to succeed in the technology market. Rather, I think companies that try to chase this emerging market at the expense of their own, established capabilities, will certainly fall behind rivals that have the good sense to work cooperatively where it makes the most sense to do so.”

IBM, SAP collaborating on large language models and generative AI

The Watson partnership with SAP is the result of a longstanding collaboration between the two companies.

Currently, SAP and IBM Consulting support customers with 25 joint intelligent industry solutions that use IBM Watson capabilities underpinned by the SAP Business Technology Platform (SAP BTP), the companies said.

“Over the past 10 or so years, the two companies have grown in such a way that they are much more complementary. But the two have been working together a lot longer than that,” Shimmin said, adding that IBM also delivers SAP BTP on top of its own cloud platform, which ties nicely into Red Hat Enterprise Linux.

In addition to embedding Watson into SAP, the companies said they are working jointly on developing large language models and generative AI capabilities targeted at delivering “consistent continuous learning and automation based on SAP’s mission-critical application suite.”