Amid key devel­op­ments in the field of gen­er­a­tive arti­fi­cial intel­li­gence (AI), in par­tic­u­lar Chat­G­PT, tech­nol­o­gy com­pa­nies of all sizes are mov­ing to both har­ness its dis­rup­tive inno­va­tion and meet the evolv­ing needs of the glob­al busi­ness com­mu­ni­ty. At the same time, cor­po­rate lead­ers must nav­i­gate mass excite­ment and con­cern from the pub­lic sec­tor about the impact of gen­er­a­tive AI across all industries. 

Dur­ing an hour-long July 11 vir­tu­al town­hall, SAP AI experts sought to strike a bal­ance between these strate­gic pri­or­i­ties. Pre­sent­ed by ASUG’s SAP S/4HANA ERP Com­mu­ni­ty Alliance, the ASUG Task Force dis­cus­sion (which you can watch in full here) explored SAP’s AI strat­e­gy and busi­ness automa­tion, with two SAP speak­ers — Raul Por­ras, Enter­prise Archi­tect, SAP, and Nagi Nal­lamil­li, Cus­tomer Offi­cer, SAP — offer­ing insights into how to imple­ment busi­ness automa­tions in SAP S/4HANA, automa­tion trends across indus­tries, and exam­ples of automa­tion use cases. 

“Arti­fi­cial intel­li­gence is a hot top­ic. It’s some­thing everybody’s talk­ing about, and there have been a lot of advances front and cen­ter for the gen­er­al pop­u­la­tion but more so for the busi­ness side,” Por­ras said, intro­duc­ing the task force. ​“The ques­tions are: How can we take advan­tage of it? How can we get ahead of it? And how do we not miss the boat?”

“Built for Business”

First reflect­ing on the recent surge of inter­est in AI tech­nol­o­gy — and on how hyper­scaler avail­abil­i­ty and the com­pute capac­i­ty to cen­tral­ize the train­ing of mod­els has fac­tored into it — Por­ras called atten­tion to the main­stream debut of AI tech­nolo­gies like Chat­G­PT and Ope­nAI. ​“It’s not that the tech­nol­o­gy has changed, but its acces­si­bil­i­ty accel­er­at­ed adop­tion and enhanced inno­va­tion,” he said. 

From doc­u­ment pro­cess­ing to rec­om­men­da­tions, fore­cast­ing, gen­er­a­tive AI, and dig­i­tal assis­tants, the poten­tial for busi­ness-spe­cif­ic AI is sig­nif­i­cant. Still, Por­ras and Nal­lamil­li don’t see it as an evo­lu­tion capa­ble of replac­ing skilled work­ers. ​“The main rea­son peo­ple are con­sid­er­ing AI is to aug­ment what humans do,” Por­ras explained.

To that end, hre said, SAP AI is ​“built for busi­ness,” being built into appli­ca­tions that already pow­er crit­i­cal busi­ness process­es, rather than requir­ing cus­tomers to change their usage of those appli­ca­tions. SAP AI will be trained with indus­try insights, busi­ness process exper­tise, and tai­lored to cus­tomers’ data, to remain rel­e­vant to dri­ving busi­ness val­ue and accel­er­at­ing busi­ness process­es. SAP AI is also built on eth­i­cal and data pri­va­cy stan­dards, ensur­ing deci­sions that come out of arti­fi­cial deci­sion-mak­ing are non-biased and respect cus­tomers’ pri­va­cy by min­i­miz­ing any amount of cus­tomer data used to retrain mod­els or pro­vide feed­back to third parties.

In dis­cussing SAP’s over­ar­ch­ing busi­ness AI strat­e­gy, Por­ras and Nal­lamil­li named cloud ERP, human cap­i­tal man­age­ment, spend man­age­ment, busi­ness net­works, and cus­tomer rela­tion­ship man­age­ment as areas in which SAP fore­sees automa­tion through AI-pow­ered busi­ness processes. 

Core to each of these is busi­ness data. Por­ras empha­sized that ​“respon­si­bly embed­ding AI capa­bil­i­ties in each” is intend­ed to help com­pa­nies uti­lize that data more effec­tive­ly. To do so, SAP plans to lever­age what exists and emerges from what it calls the ​“open ecosys­tem of gen­er­al-pur­pose AI tool­ing,” lever­ag­ing what its part­ners have con­tributed in terms of cre­at­ing, train­ing, and run­ning mod­els — includ­ing those rel­e­vant to gen­er­a­tive AI — and select­ing those providers’ tech­nolo­gies based on usage pat­terns, enter­prise-grade qual­i­ties, and oth­er such factors. 

To that effect, as far as its use of gen­er­a­tive AI goes, SAP will not use Chat­G­PT, an appli­ca­tion built on a machine learn­ing Nat­ur­al Lan­guage Pro­cess­ing mod­el, known as a Large Lan­guage Mod­el (LLM), devel­oped by Ope­nAI. But the com­pa­ny will lever­age that foun­da­tion­al LLM and its gen­er­a­tive AI capa­bil­i­ties in the con­text of busi­ness data and process­es to achieve spe­cif­ic out­comes in its appli­ca­tions, accord­ing to Porras. 

To do this ​“respon­si­bly,” SAP’s use of gen­er­a­tive AI will fol­low the same prin­ci­ples of busi­ness AI, with the same devel­op­ment and respon­si­ble AI review process­es and keep­ing humans in the loop to review and approve gen­er­at­ed infor­ma­tion. To ensure data pri­va­cy for gen­er­a­tive AI, SAP is pur­su­ing enter­prise-ready part­ner agree­ments spe­cif­ic to data pri­va­cy and iso­la­tion, to ensure no cus­tomer data is used by third-par­ty ven­dors to train foun­da­tion­al mod­els despite its usage with­in SAP appli­ca­tions to make pre­dic­tions for cus­tomers’ businesses.

“Think of AI in Layers”

Speak­ing broad­ly, Por­ras advised ASUG mem­bers to ​“think of AI in lay­ers” with­in appli­ca­tions such as SAP S/4HANA Cloud and SAP Busi­ness Tech­nol­o­gy Plat­form. The dig­i­tal core and back­bone of the S/4HANA solu­tion pro­vides a frame­work for busi­ness process­es. On top of that, intel­li­gent tech­nolo­gies such as machine learn­ing, sit­u­a­tion han­dling, and ana­lyt­ics are ful­ly embed­ded in SAP S/4HANA Cloud. Fur­ther intel­li­gent indus­try capa­bil­i­ties like intel­li­gent sit­u­a­tion automa­tion, SAP Build Process Automa­tion, and dig­i­tal assis­tants or chat­bots are also avail­able side-by-side via SAP BTP. 

With­in its focus on enrich­ing process­es and help­ing com­pa­nies make deci­sions through automa­tion, SAP is in the process of embed­ding AI capa­bil­i­ties into its exist­ing suite of appli­ca­tions. Even so, this strat­e­gy is not par­tic­u­lar­ly new to SAP. ​“Most of the appli­ca­tions that you own will have — or already have — arti­fi­cial intel­li­gence fea­tures and capa­bil­i­ties with­in them,” Por­ras said.

From pre­dic­tive data ana­lyt­ics through sta­tis­tics to mak­ing pre­dic­tions with machine learn­ing and more recent­ly cre­at­ing data mod­els to accel­er­ate busi­ness process­es with gen­er­a­tive AI, arti­fi­cial intel­li­gence func­tion­al­i­ty has evolved at SAP over the years. AI is com­pre­hen­sive already in areas such as cash appli­ca­tion, demand fore­cast­ing and sens­ing, project-cost pre­dic­tion, and sales route opti­miza­tion. ​“These fea­tures have been build­ing up since the ear­ly days of pre­dic­tive machine learn­ing,” Por­ras said. Oth­er exist­ing use cas­es for SAP AI include: 

• Intel­li­gent col­lec­tions and sales order auto-com­ple­tion for SAP S/4HANA
• Con­fig­urable prod­uct quo­ta­tion for SAP Intel­li­gent Prod­uct Recommendations
• Gen­er­al ledger line-item iden­ti­fi­ca­tion for SAP Cen­tral Invoice Management

New AI capa­bil­i­ties that will lever­age gen­er­a­tive AI, as announced dur­ing SAP Sap­phire, cur­rent­ly (or will) include: 

• Process mod­el gen­er­a­tion and doc­u­men­ta­tion for SAP Sig­navio Process Manager
• Prod­uct doc­u­men­ta­tion search for SAP Dig­i­tal Assistant
• Nat­ur­al lan­guage queries for SAP Ana­lyt­ics Cloud
• Job descrip­tion and inter­view ques­tion gen­er­a­tion for SAP SuccessFactors
• Goods receipt pro­cess­ing for SAP Trans­porta­tion Management
• Nat­ur­al lan­guage mar­ket­ing ana­lyt­ics, prod­uct descrip­tion gen­er­a­tion, and review sum­maries for SAP Dig­i­tal Assis­tant for CX

Anurag Barua, Dig­i­tal Trans­for­ma­tion Leader, SAP, not­ed ​“a lot of demand for AI and automa­tion across all my cus­tomers,” point­ing to exam­ples in finance such as ​“end-to-end automa­tion of invoice pro­cess­ing, auto­mat­ed fraud pre­ven­tion, auto­mat­ed finan­cial account rec­on­cil­i­a­tion and inter­com­pa­ny rec­on­cil­i­a­tion, detec­tion of tax com­pli­ance, auto­mat­ed cre­ation of pur­chase orders and sales orders, and using con­ver­sa­tion­al AI.” 

Barua report­ed that, on the man­u­fac­tur­ing and sup­ply chain side, he sees reduc­tion of inven­to­ry car­ry­ing costs, real-time demand fore­cast­ing, and accu­rate deliv­ery date pre­dic­tion (improv­ing sup­ply chain effi­cien­cy) as exam­ples of AI and automa­tion already hav­ing a pos­i­tive impact. ​“We are increas­ing­ly see­ing AI/​automation as a dri­ver for dig­i­tal trans­for­ma­tion across all indus­tries,” he added.

The Abil­i­ty to Adapt

As atten­dees asked ques­tions about SAP’s roadmap, var­i­ous top­ics of inter­est sur­faced, includ­ing the impor­tance of earn­ing the pub­lic sector’s trust in nav­i­gat­ing gen­er­a­tive AI at such a pro­tean stage of its devel­op­ment and focus­ing on change man­age­ment in mov­ing to explain and make acces­si­ble infor­ma­tion around the tech­nol­o­gy. Addi­tion­al­ly, one attendee reflect­ed, process man­agers embed­ding AI and automa­tion with­in exist­ing SAP solu­tions has made it chal­leng­ing for enter­prise archi­tects and tech­nol­o­gists to gain vis­i­bil­i­ty of all AI inno­va­tions across the SAP prod­uct suite.

More atten­tion must be paid, the speak­ers agreed, to out­lin­ing strate­gic objec­tives, tech­ni­cal spec­i­fi­ca­tions, and the impli­ca­tions of both for busi­ness­es, employ­ees, and the public.

Por­ras stressed the role that build­ing agili­ty with­in orga­ni­za­tions can play in prepar­ing them to con­sume AI. ​“Fore­cast accu­ra­cy is impor­tant, but the abil­i­ty to adapt after you make a fore­cast is more impor­tant,” he said, by way of exam­ple. ​“If you keep using his­tor­i­cal data or even sophis­ti­cat­ed pre­dic­tion algo­rithms, and then some­thing changes in the envi­ron­ment, that fore­cast isn’t going to help you.”

Approx­i­mate­ly 50 atten­dees — rep­re­sent­ing com­pa­nies such as Apple Inc., Bris­tol Myers Squibb, Bum­ble­bee Foods LLC., Con­stel­la­tion Brands, HP Inc., IBM, Infos­ys, John­son & John­son, Lib­er­ty Mutu­al Insur­ance, Medtron­ic, Mess­er North Amer­i­ca, Microsoft, Pacif­ic Gas & Elec­tric Com­pa­ny, Para­mount Pic­tures, Price­wa­ter­house­C­oop­ers LLC, South­ern Cal­i­for­nia Edi­son, The Nielsen Com­pa­ny, Toy­ota Motor North Amer­i­ca, Under Armour, Vitaque­st, Wal­greens, xSuite, and Zep, Inc. — reg­is­tered for and attend­ed the town hall.

More infor­ma­tion, demos, and exam­ples on busi­ness AI are avail­able from SAP. To watch the task force in full, click here.