At the begin­ning of his ses­sion at the recent Next Gen­er­a­tion SAP Enter­prise Archi­tect Learn­ing Forum, Johannes Euler, Head of AI for SAP Enter­prise Archi­tects, posed a seri­ous ques­tion to the audi­ence: Is SaaS dead? 

The audi­ence hes­i­tat­ed. A few hands went up in agree­ment. Most stayed down. 

“I like your opti­mism,” Euler joked, before refram­ing the issue. The real ques­tion isn’t around the death of SaaS but how EAs adjust to the major AI trans­for­ma­tion under­way. It’s a deep­er shift in how enter­prise sys­tems cre­ate val­ue and how EAs are now on the front line of that change. 

Thank­ful­ly, Euler out­lined five impor­tant shifts EAs need to pay atten­tion to this year to help make the AI tran­si­tion easier. 

1. From AI in Appli­ca­tions to AI on Appli­ca­tions 

For years, incre­men­tal AI fea­tures have been lay­ered into ERP and sup­ply chain sys­tems, but that mod­el is changing. 

“What we are see­ing with AI is that it is also becom­ing a com­mod­i­ty,” said Euler. ​“That leads to the shift that we should not just build AI with­in the apps, but we have an agen­tic lay­er on top of the apps.” 

This new lay­er changes every­thing. Instead of enhanc­ing indi­vid­ual appli­ca­tions, AI now orches­trates across them. For EAs, that means design­ing for AI as a con­trol plane, not just a capability. 

2. The End of Appli­ca­tion Bound­aries 

Tra­di­tion­al enter­prise archi­tec­ture has long been appli­ca­tion-cen­tric. But agen­tic AI is expos­ing the lim­its of that model. 

“When you look at sup­ply chain, it’s not just one sys­tem,” Euler said. ​“It’s maybe also some oth­er ven­dors or cus­tom solu­tions that are act­ing in this space.” 

The result is a clear shift from app bound­aries to land­scape-wide intel­li­gence, which is why orga­ni­za­tions must now archi­tect for cross-sys­tem work­flows, end-to-end busi­ness process­es, and inter­op­er­a­ble AI agents. The future isn’t about opti­miz­ing sys­tems in iso­la­tion; it’s about enabling intel­li­gence across the entire landscape. 

3. Val­ue Can’t Wait for Mod­ern­iza­tion 

For years, orga­ni­za­tions were told to mod­ern­ize before unlock­ing the val­ue of AI. But it appears that strat­e­gy might soon be outdated. 

“A lot of com­pa­nies have found a workaround,” Euler said. ​“They have found an eas­i­er way to AI that maybe isn’t as enter­prise-ready, but it some­how gets you there.” 

Even as enter­pris­es move toward plat­forms like SAP S/4HANA, busi­ness lead­ers are demand­ing imme­di­ate returns. This is cre­at­ing a new approach. 

“We have to offer you val­ue while you’re mod­ern­iz­ing and not just wait for you to be mod­ern,” said Euler. 

For archi­tects, this means design­ing par­al­lel val­ue streams, where AI deliv­ers out­comes dur­ing trans­for­ma­tion, not after it. 

4. From Human-Led to Agent-Led Trans­for­ma­tion 

Enter­prise trans­for­ma­tions have tra­di­tion­al­ly depend­ed on armies of con­sul­tants and sys­tems inte­gra­tors. That mod­el is evolv­ing fast, accord­ing to Euler. 

“Climb­ing this moun­tain is quite intense,” he said. ​“You have a lot of project effort and a lot of SI armies join­ing you.” 

Now, agen­tic AI is start­ing to take on part of that bur­den as agent-led migra­tions increase speed and qual­i­ty. The impli­ca­tion is not a replace­ment but an aug­men­ta­tion. Archi­tects must design for hybrid exe­cu­tion mod­els, where humans guide strat­e­gy and agents accel­er­ate delivery. 

5. From Ship­ping Fea­tures to Deliv­er­ing Val­ue 

One of Euler’s most strik­ing shifts had less to do with tech­nol­o­gy and more to do with oper­a­tional mindset. 

“There is a big cliff between what’s actu­al­ly there and what you are actu­al­ly using,” he said, describ­ing the gap between deliv­ered AI capa­bil­i­ties and real-world adop­tion. That gap has mul­ti­ple layers: 

• Pro­cure­ment friction 
• Acti­va­tion complexity 
• Orga­ni­za­tion­al readiness 

For EAs, suc­cess is no longer deploy­ment. It’s adop­tion and impact. 

“We have to shift away from ​‘we have shipped a capa­bil­i­ty’ to cel­e­brat­ing the actu­al val­ue real­iza­tion,” Euler said. ​“Users are actu­al­ly using it, which makes their life more productive.” 

The New Man­date for EAs 

Tak­en togeth­er, these shifts rede­fine the role itself. Enter­prise archi­tec­ture is mov­ing to the following: 

• From appli­ca­tion-cen­tric to AI- and data-centric 
• From sys­tem design to val­ue orchestration
• From gov­er­nance to enable­ment at scale 

“I think we as EAs are right in the front line to trans­late today’s world into what’s com­ing next,” said Euler, who added that this respon­si­bil­i­ty comes with a new playbook: 

• Pri­or­i­tize val­ue over capa­bil­i­ty. ​“Look always at the val­ue when design­ing use cases.” 
• Adopt before you build. ​“See what you can take off the shelf, then extend, and only then build from scratch.” 
• Pre­pare for scale ear­ly. ​“You must secure that your enter­prise is ready, not when the CIO calls you because then it’s too late.” 
• Embrace data as the foun­da­tion. ​“AI needs this con­text in order to be rel­e­vant and reliable.” 

SaaS Isn’t Dead, It’s Becom­ing Some­thing Else 

So, is SaaS dead? Not exact­ly. But it is evolv­ing into some­thing fun­da­men­tal­ly dif­fer­ent, accord­ing to Euler. 

“I decid­ed to call it work as a ser­vice and intel­li­gent work as a ser­vice,” he said. 

Under this new model: 

• Appli­ca­tions fade into the background 
• Agents orches­trate work across systems 
• Val­ue is deliv­ered through out­comes, not interfaces 

For EAs, the future isn’t about design­ing appli­ca­tions. It’s about design­ing how work hap­pens in a world where agents are increas­ing­ly doing the work.