SAP recent­ly announced an expan­sion of its part­ner­ship with NVIDIA focused on ​“fine-tun­ing” large lan­guage mod­els (LLMs), set­ting the stage for its cus­tomers to begin adopt­ing gen­er­a­tive AI across the SAP suite of solutions. 

The announce­ment, made on March 18, detailed plans to embed NVIDIA’s gen­er­a­tive AI foundry ser­vice into SAP Data­s­phere, SAP Busi­ness Tech­nol­o­gy Plat­form (BTP), RISE with SAP, and addi­tion­al SAP enter­prise solu­tions, as well as to accel­er­ate appli­ca­tion deploy­ment through NVIDIA’s NIM microser­vices platform. 

“Through our recent col­lab­o­ra­tion with NVIDIA, we accel­er­ate gen­er­a­tive AI adop­tion across our appli­ca­tion and solu­tion port­fo­lio, increas­ing the lev­el of val­ue for our cus­tomers around the globe,” Philipp Herzig, Chief AI Offi­cer at SAP told ASUG. ​“Togeth­er, we are rev­o­lu­tion­iz­ing the way our cus­tomers expe­ri­ence AI, unlock­ing new poten­tial for their success.” 

Build­ing upon SAP CEO Chris­t­ian Klein’s ear­li­er stat­ed inten­tion to trans­form SAP into the ​“no. 1 busi­ness AI com­pa­ny,” the new capa­bil­i­ties aim to help SAP cus­tomers begin lever­ag­ing gen­er­a­tive AI solu­tions across the entire SAP suite of cloud solu­tions and appli­ca­tions. The part­ner­ship includes the fol­low­ing initiatives: 

• Retrieval-aug­ment­ed gen­er­a­tion capa­bil­i­ties (RAG) for Joule, SAP’s gen­er­a­tive AI copi­lot. This AI-opti­miza­tion process trains LLMs, extend­ing their knowl­edge bases to cov­er spe­cif­ic ver­ti­cals while ensur­ing accu­ra­cy and relevance. 
• The cre­ation of more than 20 gen­er­a­tive AI use cas­es for SAP S/4HANA Cloud, SAP Suc­cess­Fac­tors, and SAP Sig­navio. Exam­ples of use cas­es the com­pa­nies are pur­su­ing include intel­li­gent invoice match­ing in SAP S/4HANA Cloud, and improved busi­ness rec­om­men­da­tions and cus­tomer sup­port process­es in SAP Signavio. 
• LLM capa­bil­i­ties for SAP’s ABAP pro­gram­ming lan­guage to help devel­op­ers cre­ate domain-spe­cif­ic lan­guage code. 
• The abil­i­ty to access AI and ML (machine learn­ing) mod­els with SAP Data­s­phere, enabling uni­fied, cohe­sive data sources from SAP and third-par­ty applications. 

SAP and NVIDIA said these new capa­bil­i­ties will be avail­able by the end of 2024.

Fur­ther­more, SAP will enhance the SAP Busi­ness Tech­nol­o­gy Plat­form (BTP) with gen­er­a­tive AI capa­bil­i­ties pow­ered by NVIDIA DGX Cloud AI super­com­put­ing, NVIDIA AI Enter­prise soft­ware, and NVIDIA AI Foun­da­tion mod­els, fur­ther enabling cus­tomers to make use of gen­er­a­tive AI in their SAP environments. 

The news comes just months after SAP dou­bled down on its Busi­ness AI direc­tion, announc­ing a €2 bil­lion restruc­tur­ing pro­gram aimed at reskilling SAP employ­ees in AI, and invest­ing approx­i­mate­ly €1 bil­lion in AI use cas­es for cus­tomers. Addi­tion­al­ly, SAP recent­ly unveiled gen­er­a­tive AI and data gov­er­nance capa­bil­i­ties for SAP Data­s­phere and SAP Ana­lyt­ics Cloud, empha­siz­ing the impor­tance that a holis­tic data and ana­lyt­ics strat­e­gy will play in ensur­ing that SAP cus­tomers can effec­tive­ly lever­age AI in their tech­nol­o­gy landscapes. 

“Enter­prise cus­tomers want to lever­age state-of-the-art tech­nol­o­gy that deliv­ers real busi­ness val­ue,” said Chris­t­ian Klein, CEO and Mem­ber of the Exec­u­tive Board of SAP SE, in an offi­cial state­ment. ​“Strate­gic tech­nol­o­gy part­ner­ships, like the one between SAP and NVIDIA, are at the core of our strat­e­gy to invest in tech­nol­o­gy that max­i­mizes the poten­tial and oppor­tu­ni­ty of AI for busi­ness. NVIDIA’s exper­tise in deliv­er­ing AI capa­bil­i­ties at scale will help SAP accel­er­ate the pace of trans­for­ma­tion and bet­ter serve our cus­tomers in the cloud.”

“SAP is sit­ting on a gold mine of enter­prise data that can be trans­formed into cus­tom gen­er­a­tive AI agents to help cus­tomers auto­mate their busi­ness­es,” added Jensen Huang, founder and CEO of NVIDIA, in an offi­cial state­ment. ​“Togeth­er, NVIDIA and SAP will bring cus­tom gen­er­a­tive AI to the thou­sands of enter­pris­es around the world that rely on SAP to pow­er their operations.”

Data Qual­i­ty Remains Top of Mind

In the SAP ana­lyst com­mu­ni­ty, news of the expand­ed NVIDIA part­ner­ship was received with a bal­ance of inter­est and caution. 

Accord­ing to Joshua Green­baum, Prin­ci­pal at Enter­prise Appli­ca­tions Con­sult­ing (EAC), SAP cus­tomers could even­tu­al­ly ben­e­fit from learn­ing more about NVIDIA’s microser­vices, but their more imme­di­ate focus is on the bar­ri­ers that data qual­i­ty and data gov­er­nance pose to effec­tive­ly lever­ag­ing gen­er­a­tive AI at scale. 

“For the SAP cus­tomer base, the biggest chal­lenge for suc­cess­ful deploy­ment of advanced AI that can solve the real-world prob­lems of sup­ply chain con­straints, logis­tics bot­tle­necks, per­son­nel deploy­ment require­ments, asset man­age­ment com­plex­i­ty, and the like remains the uni­ver­sal prob­lem of data qual­i­ty and gov­er­nance,” he said. ​“Build­ing suc­cess­ful AI mod­els of any size or func­tion­al com­port­ment requires impec­ca­bly clean and his­tor­i­cal­ly valid data.” 

Green­baum added that this col­lab­o­ra­tion could ben­e­fit SAP cus­tomers in the future, once they’ve addressed more press­ing data issues.