When you name your com­pa­ny after a Ger­man phrase that trans­lates to ​“sun­ny weath­er,” it’s an indi­ca­tion of an opti­mistic out­look. That for­ward-look­ing approach is what dri­ves inno­va­tion at Kaiser­wet­ter Ener­gy Asset Man­age­ment, a glob­al renew­able ener­gy com­pa­ny that deliv­ers spe­cial­ized ser­vices to help invest­ment firms, banks, and gov­ern­ments have vis­i­bil­i­ty into their invest­ments in solar parks and wind farms around the world. 

SAP Leonar­do Intel­li­gence Inspires ARISTOTELES

As it was build­ing its cloud-based Inter­net of Things (IoT) plat­forms, Kaiser­wet­ter turned to SAP intel­li­gent tech­nolo­gies for the frame­work it uses to sup­port its core prod­uct, known as ARIS­TOTE­LES. Rather than cre­at­ing these inno­va­tions from scratch, Kaiser­wet­ter opt­ed to com­bine smart data ana­lyt­ics, pre­dic­tive ana­lyt­ics, and machine learn­ing from SAP into what it describes as a data-ana­lyt­ics-as-a-ser­vice prod­uct (DAaaS).

We had the oppor­tu­ni­ty to speak with Kaiser­wet­ter CEO and Founder, Han­no Schok­l­itsch, to hear about how the ARIS­TOTE­LES plat­form inter­con­nects tech­nolo­gies such as arti­fi­cial intel­li­gence (AI), machine learn­ing, IoT, and pre­dic­tive ana­lyt­ics to help its cus­tomers max­i­mize their invest­ment returns while min­i­miz­ing risk. 

Ann Marie: Could you tell us a bit about what Kaiser­wet­ter does? Who are your typ­i­cal cus­tomers and who are the end users of your dashboards?

Han­no: Our end users are investors and invest­ment funds. We’ve been doing this work to cat­alyze invest­ment into renew­able ener­gy. As we all know, there’s a need for a lot of cap­i­tal to sup­port putting more renew­able ener­gy pro­duc­tion in place so that we can achieve a zero-emis­sion future. 

On the oth­er side, we also are talk­ing to lend­ing banks about unlock­ing the cap­i­tal for financ­ing renew­able ener­gy. The idea is to under­stand how we can real­ly max­i­mize returns of the cash flow from these invest­ments while min­i­miz­ing risk and offer­ing trans­paren­cy to our stake­hold­ers. This is exact­ly what we are bring­ing togeth­er from all the data and ana­lyt­ics in our prod­ucts and what we pro­vide to our clients through our dashboards. 

Ann Marie: What types of projects are these banks funding? 

Han­no: We are fund­ing wind parks and solar parks. And we are now work­ing on inte­grat­ing our first hydropow­er sta­tion into our ARIS­TOTE­LES plat­form — of course, using all the great ser­vices that SAP is pro­vid­ing us to get our data in the right place and run our analytics. 

Ann Marie: Tell me a lit­tle bit more about what your ARIS­TOTE­LES plat­form does and how it’s dif­fer­ent from any­thing else on the market.

Han­no: One of the essen­tial dif­fer­ences is that we have data as a ser­vice as our busi­ness angle. We are not talk­ing about soft­ware as a ser­vice. We are gath­er­ing the data for our clients, adding ana­lyt­ics, and giv­ing it back to them in a for­mat they can use to make decisions. 

Every­body knows that there’s a lot of data fly­ing around. It’s not about just col­lect­ing the data but get­ting the right sup­port in place to draw the intel­li­gence out. This is exact­ly what we are offer­ing to our clients. Our clients achieve data trans­paren­cy. Using SAP Data Intel­li­gence, we are find­ing insights and deliv­er­ing finan­cials to our dashboards. 

For the investors who are our clients, we must rein­force how their cap­i­tal is per­form­ing and whether any issues may arise from a par­tic­u­lar site that could affect their capital. 

Ann Marie: How does the plat­form help your cus­tomers make deci­sions? You men­tioned that it pro­vides infor­ma­tion to help them decide not only to make the invest­ment in the first place, but also to man­age risk once they do. 

Han­no: When we are ana­lyz­ing spe­cif­ic machines, we use mod­el­ing. For exam­ple, each of the wind tur­bines at the wind farm has a bar­code, which is deliv­ered by the man­u­fac­tur­er. In some cas­es, we mod­el what this wind tur­bine will do — even though it is not data com­ing from the actu­al turbine. 

So, it’s not just data col­lec­tion or some­thing like busi­ness intel­li­gence. It’s get­ting the data from the asset, and that data speaks to the mod­el­ing data. Then we get the data into a bar curve and watch for where any devi­a­tion is com­ing from. That gives us the right base­line in a smart data ana­lyt­ics approach.

Ann Marie: How does SAP tech­nol­o­gy come into the picture? 

Han­no: We are apply­ing AI and SAP Data Intel­li­gence to put machine learn­ing algo­rithms in place while using the capa­bil­i­ties of com­put­ing pow­er from SAP to run our models.

This gives our clients the abil­i­ty to see if there is a poten­tial trend hap­pen­ing and to iden­ti­fy this at an ear­ly stage, which is a big, big advan­tage for them. We put our machine learn­ing in the front end of our prod­uct. Our clients are open­ing their note­books or desk­tops in the morn­ing and they can see — What is the machine learn­ing algo­rithm telling me? What is the fore­cast of my ener­gy production? 

They can see the stan­dard devi­a­tion between the ana­lyt­ics and the real pro­duc­tion. If this devi­a­tion is increas­ing, then you poten­tial­ly have already iden­ti­fied an issue com­ing up. They can mon­i­tor the first stage to under­stand the per­for­mance of their assets and whether there are any prob­lems with the performance. 

This is a great step to not just have data and make nice and shiny charts out of it. In the end, what we are offer­ing is much more than a report­ing setup.

Ann Marie: You are using SAP Leonar­do for AI, data intel­li­gence, and IoT ser­vices. What devices are you con­nect­ing and mon­i­tor­ing through IoT? 

Han­no: We are using an IoT set­up and putting the con­nec­tors into the SAP Cloud Platform using Dock­er con­tain­ers, which are then retriev­ing the data from each of the assets. This is one of the points where we are con­nect­ing direct­ly with the assets at wind farms and solar parks all around the world. We are col­lect­ing the data that is com­ing out of each oper­a­tion and using soft­ware on-site to gath­er the data, which is stored for 10 min­utes and sent to us. This is suf­fi­cient for our ana­lyt­ics to do the right cor­re­la­tions between the data. 

Ann Marie: How do you inte­grate pre­dic­tive ana­lyt­ics into the process?

Han­no: This is what we use machine learn­ing for. We can pre­dict what ener­gy or pow­er the asset should be cre­at­ing with­in the next few hours. And then we have the real data com­ing from the asset. If this does not match or shows an increase in the devi­a­tion, you can already see that there is an issue com­ing up. Based on these num­bers, you can set mea­sures to avoid any poten­tial threat or neg­a­tive impact. 

We have returns of 95 – 97%, which is great per­for­mance that we can show to our clients. This is pos­si­ble because we are using the machine learn­ing algo­rithms we have put in place and because of our com­put­ing pow­er, which we get from SAP. 

Ann Marie: How do you build your data models?

Han­no: We start with pre­fab­ri­cat­ed mod­els from SAP, which, of course, must be cus­tomized to meet our needs. We have our own data sci­en­tists who adjust these pre­fab­ri­cat­ed mod­els. If we don’t have what we need, we also make our own cus­tomized mod­els. I think the pre­fab­ri­cat­ed mod­els are help­ful, but you real­ly need your own data sci­en­tist team. 

Ann Marie: How do you incor­po­rate weath­er data into your analysis?

Han­no: We have third-par­ty weath­er data that we are inte­grat­ing for the wind and solar parks. This is not just a weath­er mod­el itself — we are actu­al­ly get­ting the pow­er fore­cast based on the weath­er mod­els. We look at the pow­er the machines are expect­ed to pro­duce for the next 24 hours, and then for three to sev­en days. 

Ann Marie: Why did you choose to rely on the cloud for your solution?

Han­no: I think it’s the only way for­ward to han­dle all this big data stuff. We already have so much data that we see no oth­er pos­si­bil­i­ty than using the cloud. We are a small­er com­pa­ny — but even for medi­um-sized com­pa­nies — I think the future must be run by the cloud. The cloud was crit­i­cal to get the high-per­for­mance data man­age­ment we need­ed, as well as hav­ing SAP HANA in place for its in-mem­o­ry, high-speed data ana­lyt­ics. This was, to us, essen­tial to our operations. 

Ann Marie: I would assume that hav­ing a cloud-based solu­tion makes your offer­ing more acces­si­ble to your cus­tomers, as well. 

Han­no: Absolute­ly. We have so many pos­si­bil­i­ties now. For exam­ple, we have cus­tomers who do the data ana­lyt­ics with an appli­ca­tion on-site and they may be in a loca­tion with extra-cloudy weath­er. They are con­nect­ing edge servers to the SAP Cloud Plat­form to get the data and ana­lyt­ics on our platform. 

We are see­ing an increas­ing world of dif­fer­ent setups out there, which we have to adopt quick­ly into our plat­form. The SAP Cloud Plat­form allows us to adapt quite fast to these ongo­ing changes in dif­fer­ent data inte­gra­tions with assets con­nect­ed by IoT. We are always look­ing to con­nect faster and bet­ter. Data qual­i­ty is essen­tial to us, and we are work­ing to improve data integri­ty. As we bring our tech­nol­o­gy to dif­fer­ent areas in the world, we are thank­ful to have the use of SAP capabilities. 

Ann Marie: What advice would you offer oth­er SAP cus­tomers who are look­ing to use AI and machine learning?

Han­no: I am an entre­pre­neur. When we tran­si­tioned to a dig­i­tal com­pa­ny, for me it was cru­cial not to build our own plat­form com­plete­ly by our­selves. I’m con­vinced that it’s much more dili­gent to use the plat­forms that tech­nol­o­gy com­pa­nies are already pro­vid­ing that can be cus­tomized to your own busi­ness model. 

We are build­ing a busi­ness mod­el that relies on a tech­nol­o­gy plat­form that SAP is pro­vid­ing using AI, IoT, and SAP Ana­lyt­ics Cloud. The whole prod­uct suite of SAP is quite big. And work­ing with SAP is great because you get com­put­ing pow­er. If you would like to per­form AI on your own, you will ulti­mate­ly be lim­it­ed in the end by your com­put­ing pow­er. Get­ting the com­put­ing pow­er to do this is a big advantage. 

Final­ly, you need to have data sci­en­tists on your side to work with AI and machine learn­ing. It’s not that easy. Even with the pre­fab mod­els, you need data sci­en­tists who under­stand your busi­ness mod­el and who can real­ly focus on how you can sup­port that mod­el using AI. 

Ann Marie: Over the long term, how are your solu­tions help­ing insti­tu­tions invest more sustainably?

Han­no: If you’re going, for exam­ple, to invest in these kinds of assets in oth­er coun­tries — even in the U.S. or Europe or South Amer­i­ca — you are fac­ing a poten­tial­ly high­er risk/​return pro­file. You have to watch what is going on, and that is what we’re doing. The tech­ni­cal part of these invest­ments, we have com­plete­ly under con­trol. We know exact­ly what the machine is telling us and what ener­gy it is pro­duc­ing. We have the finan­cials under con­trol. We know where the accounts are, as we are con­nect­ed with the banks and we know where the cash is. In the end, this is high­ly effi­cient risk man­age­ment. What we have today, we wouldn’t have had the tech­nol­o­gy to do two or three years ago. Now, you can invest in dif­fer­ent mar­kets, even if the risk is high­er, because we can con­trol that risk in a dif­fer­ent way.

We can also turn the world around on the financ­ing of solar and wind farms, or oth­er sim­i­lar projects, by talk­ing to the lend­ing banks. How are they cur­rent­ly watch­ing their credit/​loan port­fo­lios? I think this must change com­plete­ly. Cur­rent­ly, banks are look­ing into the past. They are grant­i­ng a loan to a cus­tomer for a wind farm or solar park. And then the cus­tomer has the oblig­a­tion to send them reports on a month­ly basis. All of these reports are look­ing into the past. We have to change this — we have to look into the future. 

And this is exact­ly what Kaiser­wet­ter already pro­vides. You can look into the future and see if there’s a poten­tial issue com­ing up. Dis­tressed assets should not sud­den­ly become a prob­lem for banks if they are using data to pre­dict the per­for­mance of these assets in the future. 

As for the next pos­si­bil­i­ty? Infor­mat­ics is here. What if you opened your desk­top or lap­top and it start­ed talk­ing to you? We have iden­ti­fied which of your invest­ments will have poten­tial threats com­ing in three or four months. This is the future — the machine is talk­ing to you and you can act. This type of data could help dri­ve your deci­sion-mak­ing process. 

Ann Marie: Han­no, thank you so much for your time today, and for explain­ing how you sup­port investors in renew­able energy. 

Han­no: Call me when­ev­er you want to have good weath­er. You know, Kaiser­wet­ter is known for mak­ing good weather.

Don’t miss hear­ing about oth­er orga­ni­za­tions using intel­li­gent tech­nolo­gies, includ­ing those in the ener­gy and util­i­ties space, at our vir­tu­al expe­ri­ence ASUG Best Prac­tices: SAP for Indus­tries in Sep­tem­ber. Reg­is­ter today.