Evaluating
Digital Maturity

Intro­duc­ing our expert: Kevin Lin is a fas­ci­nat­ing guide for any­one wish­ing to explore the world of arti­fi­cial intel­li­gence. He can whisk peo­ple away on an excit­ing vir­tu­al jour­ney into the dig­i­tal life of the future. The voy­age is as cap­ti­vat­ing as a night at the movies. But sci­ence fic­tion is not his thing. Lin’s exam­ples of AI appli­ca­tions are real and reflect main­stream busi­ness life. He takes a mat­ter-of-fact approach and seeks con­crete solu­tions that truly help com­pa­nies move for­ward. And he con­sis­tent­ly main­tains a crit­i­cal dis­tance to the hype about an El Dora­do of data.

Born in Tai­wan in 1978, Lin stud­ied mechan­i­cal engi­neer­ing and semi­con­duc­tor tech­nol­o­gy in his home­land before com­ing to Ger­many to com­plete his engi­neer­ing stud­ies at the pres­ti­gious RWTH Aachen Uni­ver­si­ty in North Rhine-West­phalia with a focus on avi­a­tion and aero­nau­tics. A 2004 trainee­ship in sus­pen­sion devel­op­ment brought him into con­tact with Porsche AG for the first time. Short­ly there­after he began work­ing as a test engi­neer at the Stuttgart-based sports-car maker’s devel­op­ment cen­ter in Weissach.

While pur­su­ing his doc­tor­ate at the Tech­ni­cal Uni­ver­si­ty of Darm­stadt, Lin worked inten­sive­ly on AI-sup­port­ed causal analy­ses of busi­ness strate­gies and deci­sions. And in 2015 he joined Porsche Con­sult­ing. Now the tech­ni­cal direc­tor for a ten-mem­ber expert team, he is respon­si­ble for the man­age­ment consultancy’s Arti­fi­cial Intel­li­gence and Data Ana­lyt­ics field togeth­er with Asso­ciate Part­ner Fabi­an Schmidt. In his per­son­al life—perhaps as a coun­ter­bal­ance to his work—Lin has a fond­ness for ana­logue tech­nolo­gies. With great atten­tion to detail, he is restor­ing a Beetle—or to be pre­cise, a 1973 Volk­swa­gen 1303 S with 50 horse­pow­er, a 1.6‑liter capac­i­ty, and a Raven­na Green paint job. “It’s near­ly fin­ished,” he says. “Now all that’s need­ed is a com­plete over­haul of the orig­i­nal interior.”

In this inter­view Kevin Lin dis­cuss­es whether busi­ness and indus­try are pay­ing suf­fi­cient atten­tion to the poten­tial of arti­fi­cial intel­li­gence, and how data ana­lyt­ics can help them succeed.

“On the one hand, we’re work­ing with ever greater vol­umes of data. Yet on the other, we’re expect­ing ever more detailed answers to com­plex ques­tions. Excel spread­sheets and sim­ple math­e­mat­ics are not up to the job when it comes to things like spot­ting trends or iden­ti­fy­ing sources of error in com­plex con­texts, for exam­ple. For that we need dig­i­tal tools like machine learn­ing and arti­fi­cial intel­li­gence. Those who don’t mas­ter the use of dig­i­tal tools will find it hard to inter­pret the phys­i­cal world in the future. The future via­bil­i­ty of a com­pa­ny or orga­ni­za­tion will depend on how effi­cient­ly it is able to use this set of tools. I’m con­vinced of that.”

“No, not at all. AI and data ana­lyt­ics appli­ca­tions are still in the very early stages, espe­cial­ly in Ger­many. As I see it, there are three major rea­sons for this. First of all, a lot of com­pa­nies aren’t focus­ing on the right issues. Many are eager to put arti­fi­cial intel­li­gence on their agen­das, but in actu­al prac­tice they often over­look the need to lay solid foun­da­tions for it. By that I mean cre­at­ing the basic pre­req­ui­sites like the infra­struc­ture need­ed for AI appli­ca­tions. Sec­ond, the qual­i­ty of their exist­ing data is fre­quent­ly not good enough. Vol­ume alone doesn’t make data valu­able, even though that’s a com­mon assump­tion. More­over, many com­pa­nies don’t define the respon­si­bil­i­ties of their data own­ers, mean­ing the peo­ple in charge of data man­age­ment and qual­i­ty, in clear and uni­form terms. Effec­tive man­age­ment would be help­ful here, with over­ar­ch­ing data gov­er­nance. It sets guide­lines, checks adher­ence, and there­by ensures that the oper­a­tions ben­e­fit. The third fac­tor act­ing as a brake on progress has to do with a high­ly sen­si­tive and com­pli­cat­ed mat­ter, name­ly the way in which data pro­tec­tion is under­stood. Legal con­sid­er­a­tions often pre­vent avail­able infor­ma­tion from being used. It’s like hav­ing a buried trea­sure but not being allowed to dig it up.”

“Retail is out in front. Com­par­a­tive­ly speak­ing, sales and logis­tics start­ed using dig­i­tal sys­tems very early on—in e‑commerce, for instance. The data in this area allow us to gen­er­ate trans­paren­cy and rec­og­nize pat­terns. By con­trast, in pub­lic sec­tors like trans­porta­tion, reg­u­la­tions or per­haps also a reluc­tance to invest are hold­ing back the intro­duc­tion of AI and data ana­lyt­ics solu­tions. Impor­tant use data are often not even com­piled, even though other coun­tries with sim­i­lar data pro­tec­tion require­ments have long been devel­op­ing prac­ti­cal solu­tions. That means the foun­da­tions aren’t there for use-based main­te­nance or for infra­struc­ture expan­sion plans.”

“It depends. The amount of effort need­ed for a given com­pa­ny is a func­tion of what’s known as its dig­i­tal matu­ri­ty. That has to be eval­u­at­ed. As man­age­ment con­sul­tants, we there­fore start by check­ing whether viable foun­da­tions are already part of the company’s cul­ture. We look at key fac­tors like data qual­i­ty and data literacy—meaning the abil­i­ty to eval­u­ate data and uti­lize them in thought-out ways. Anoth­er fac­tor is the added value that AI-sup­port­ed deci­sions can give the busi­ness. That’s the basis for decid­ing how much to invest in devel­op­ing AI solu­tions that will ben­e­fit the bot­tom line. Spe­cial­ists known as data sci­en­tists are also need­ed. They pos­sess the cru­cial abil­i­ty to com­bine human cre­ativ­i­ty and AI ser­vices in ways that achieve the best pos­si­ble out­put. That’s why there’s been so much com­pe­ti­tion for tal­ent­ed AI peo­ple over the years, with tech giants in par­tic­u­lar offer­ing exor­bi­tant salaries.”

“New tech­nolo­gies have always meant hav­ing to learn and relearn. Right now what’s impor­tant is to under­stand what pos­si­bil­i­ties AI and data ana­lyt­ics can open up. And to learn how to use these tools as suc­cess­ful­ly as pos­si­ble. There’s no doubt that when­ev­er you intro­duce new things into life, and espe­cial­ly at work­places, there will almost always be an ele­ment of fear or anx­i­ety. We con­sul­tants take these gen­er­al­ly fore­see­able reac­tions very seri­ous­ly. In our expe­ri­ence, the best way to encour­age accep­tance in this type of project is for every­one affect­ed by the change to con­tribute their spe­cial­ized knowl­edge and help shape the tran­si­tion, start­ing with the AI mod­el­ling phase. AI mod­els that have been assessed by spe­cial­ists from dif­fer­ent dis­ci­plines yield qual­i­ta­tive­ly bet­ter deci­sions. And the employ­ees all gain the sense that they still have their hands on the wheel. They then per­ceive AI as a help­ful tool that can be of ben­e­fit to them in their indi­vid­ual jobs. There’s a lot of room to help shape this process. Because one thing is clear: the role that peo­ple play in intro­duc­ing arti­fi­cial intel­li­gence is ulti­mate­ly deter­mined by the peo­ple themselves.”

More about arti­fi­cial intel­li­gence: Ten Hur­dles Slow­ing the Transformation 

Direct con­tact with Porsche Con­sult­ing, Arti­fi­cial Intel­li­gence and Data Ana­lyt­ics depart­ment: ai@porsche-consulting.com